JP4208288B2 - Search menu creation device, search menu creation method, and recording medium recording search menu creation program - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a search menu creation device that creates a search menu corresponding to the content of data from a plurality of data, a search menu creation method, and a recording medium on which a search menu creation program is recorded.
[0002]
[Prior art]
In recent years, a large amount of data can be easily obtained by an individual via a network such as the Internet. A menu search system that creates a search menu according to the contents of the data so that even a user unfamiliar with information equipment can easily obtain desired information from the large amount of data thus obtained is an information processing society. It is introduced in Research Report 97-NL-117-14 “Automatic Classification of Document Data Using Thesaurus”. FIG. 41 is a block diagram showing the configuration of this menu search system.
[0003]
The menu search system shown in FIG. 41 includes a data storage unit 3501, a thesaurus storage unit 3502, a data feature point extraction unit 3503, a data association unit 3504, a menu creation unit 3505, and a display unit 3506.
A data storage unit 3501 stores document data acquired via a network or the like and feature points extracted by the data feature point extraction unit 3503. FIG. 42A shows an example of data stored in the data storage unit 3501. As shown in this figure, the data storage unit 3501 stores document data 3621, 3622, 3623 acquired via a network and the like, and feature points 3601, 3602, 3603 extracted by the data feature point extraction unit 3503. is doing.
The thesaurus storage unit 3502 stores a thesaurus. The thesaurus is a dictionary in which words are classified and arranged in a tree structure according to meaning. The thesaurus storage unit 3502 arranges higher concept words closer to the root of the tree structure in this thesaurus, and arranges lower concept words closer to the leaves of the tree structure. Arrange.
[0004]
FIG. 42B shows a thesaurus 3511 as an example of the thesaurus stored in the thesaurus storage unit 3502. In this thesaurus 3511, “transportation” 3618 is the root, “transportation” 3618 is connected to “car” 3617 and “railway” 3616, and “car” 3617 is connected to “truck” 3611 and “traffic”. A “bus” 3612 is connected, and a “railway” 3616 is connected with a “train” 3613 and a “train” 3614.
In this thesaurus 3511, “transportation” 3618 is a superordinate concept of “car” 3617 and “railway” 3616, “car” 3617 is a superordinate concept of “truck” 3611 and “bus” 3612, and “railway” "3616" is a superordinate concept of "train" 3613 and "train" 3614.
[0005]
A data feature point extraction unit 3503 extracts feature points indicating features of the document data from the document data stored in the data storage unit 3501. Specifically, words that frequently appear in the document data are extracted as important words and used as feature points. In FIG. 42A, feature points 3601 “bus”, feature points 3602 “truck”, and feature points 3603 “train” are feature points extracted by the data feature point extraction unit 3503 from the document data 3621, 3622, and 3623, respectively. It is.
The data association unit 3504 associates the feature points with the words in the thesaurus having the same expression as the feature points. For example, the data association unit 3504 associates the feature point 3601 “bus” with the “bus” 3612 in the thesaurus 3511 having the same expression as the feature point 3601 “bus”. Similarly, the data association unit 3504 associates the feature point 3602 “track” and the feature point 3603 “train” with “track” 3611 and “train” 3614 in the thesaurus 3511, respectively. In this way, as shown in FIG. 43, a thesaurus 3511a associated with the feature points is generated.
[0006]
The menu creation unit 3505 extracts a word associated with a feature point and a word connected to the word from the thesaurus 3511a that has been associated with data, and generates a menu structure. In the thesaurus 3511a shown in FIG. 43, the menu creation unit 3505 includes a “track” 3611 associated with the feature point 3602 “track”, a “bus” 3612 associated with the feature point 3601 “bus”, and a feature point. “Train” 3614 associated with 3603 “Train”; “Car” 3617 connected to “Track” 3611; “Railway” 3616 connected to “Train” 3614; “Transport” connected to “Car” 3617 The “engine” 3618 is extracted to generate a menu structure 3511b shown in FIG.
[0007]
Display unit 3506 displays a search menu according to the generated menu structure. Based on the menu structure 3511b shown in FIG. 44 (a), the display unit 3506 has an option “car” 3702 and an option “railway” 3703 with the title “transportation” 3701 as shown in FIG. 44 (b). A search menu 3700 consisting of Here, when the user selects the option “car” 3702, the display unit 3506 next displays the option “truck” 3712 and the option “bus” 3713 with “car” 3711 as the title, as shown in FIG. A search menu 3710 consisting of is displayed.
[0008]
[Problems to be solved by the invention]
However, in the above conventional example, since a search menu using a thesaurus is generated based on data acquired via a network or the like, even a user unfamiliar with information equipment can easily obtain desired information. Although there is an advantage that a search menu that can search for is generated, the search menu is generated using one thesaurus, so the search target data does not apply to a thesaurus in which words are classified and arranged in a tree structure by meaning There is a problem that there are things. For example, the word “road” is related to “transportation”, but does not correspond well to the words in the thesaurus 3511 classified and arranged from the viewpoint of “transportation”.
[0009]
In addition, an age menu having options “under 10 years old”, “11-20 years old”, “21 years old or older”, and options “student”, “company employee”, “ The search menu consisting of the occupation menu with "Housewife" cannot be generated based on the thesaurus because the options included in the age menu and occupation menu have different viewpoints of classification and are not connected in that sense. There is a point.
[0010]
Furthermore, when adding data to the thesaurus or changing the data of the thesaurus, it is necessary to maintain the structure specific to the thesaurus that the data is categorized and arranged in a tree structure in the meaning. There is a problem that the operation of adding or changing is very complicated.
The present invention relates to a search menu creation device, a search menu creation method, and a search menu creation for generating a search menu by properly associating data with respect to options included in classification patterns that have different viewpoints of classification and are not connected in meaning. An object is to provide a recording medium on which a program is recorded.
[0011]
[Means for Solving the Problems]
  In order to solve this problem, the present inventionA search menu creation device for creating a search menu including options used for searching for a plurality of input data, a data storage means having a storage area for storing a plurality of input data, and a plurality of search menu options Classification pattern storage means in which a plurality of classification patterns composed of the above items are stored in advance, receiving means for receiving a plurality of input data, writing means for writing the received plurality of input data to the data storage means, Association means for associating each of the plurality of input data written in the data storage means with an item included in any of the plurality of classification patterns stored in advance in the classification pattern storage means; and the classification pattern Each item constituting the classification pattern for each of a plurality of classification patterns stored in advance in the storage means Item that constitutes the selected classification pattern, and the classification pattern selection means for selecting the classification pattern based on the variance of the input data calculated from the counted value Menu extraction means for extracting an item associated with the input data, generating a search menu using the extracted item as an option, and display means for displaying the search menu generated by the menu extraction means;It is characterized by providing.
[0014]
Here, the data storage means stores a data identifier for identifying each input data, and the associating means sequentially reads the input data from the data storage means, and inputs the input data for each read input data. A feature point extraction unit for extracting feature points included in the data; an item search unit for searching for an item corresponding to the feature point from the classification pattern storage unit for each of the extracted feature points; and the searched item In addition, a data association unit for adding a data identifier for identifying the input data may be included.
[0015]
Here, each input data stored in the data storage means is document data, and the feature point extraction unit includes a data reading element for sequentially reading the document data from the data storage means, and the read document data A morphological analysis element that performs morpheme analysis on each document data and decomposes it into morphemes for each document data, and extracts important words from the decomposed morphemes, and important words having a part or all of the extracted important words as feature points You may comprise so that an extraction element may be included.
[0016]
Here, the important word extraction element is an important word candidate extraction element that extracts a keyword candidate from the decomposed morpheme, and the number of times that the same keyword candidate as the extracted keyword candidate appears in the same document data. The number of appearance count elements for counting and the extracted important word candidates are extracted as important words from the extracted important word candidates, and the extracted number of important words is a predetermined number of times. You may comprise so that an extraction element may be included.
[0017]
Here, the important word extraction element is a key word candidate extraction element that extracts a key word candidate from the decomposed morpheme, and the number of times that the same key word candidate as the extracted key word candidate appears in the same document data. A predetermined number of important word candidates are extracted as important words in order from the extracted important word candidates and the extracted important word candidates in order from the important word candidates with the largest number of counted times. It may be configured to include a predetermined number of extraction elements having a feature point.
[0018]
Here, the data storage means stores the plurality of input data as a relational database, each input data corresponds to a row of the relational database, and is composed of an array of at least one value. A data reading element for sequentially reading input data from the data storage means and a data extracting element for extracting at least one value constituting the read input data as a feature point may be included.
[0019]
Here, the items included in the classification pattern storage unit form a hierarchical structure, and the association unit further extracts, for each extracted feature point, the item corresponding to the feature point as the classification pattern. A hierarchical item search unit that searches from storage means may be included, and the data association unit may be configured to associate the searched item with the input data including the feature point.
[0020]
Here, further, an input receiving unit that accepts selection of an option from the displayed search menu from the user, and when the input receiving unit receives the user's selection, a search menu is created based on a predetermined criterion. Determining means for determining whether or not to display input data, and, if the determination means determines that a search menu is to be created, from the items subordinate to the item corresponding to the option for which the selection is accepted An item menu generating unit that extracts an item associated with input data and generates a search menu using the extracted item as an option; and the determination unit generates the search menu when it is determined to generate the search menu. If the search menu is displayed and it is determined that the input data is displayed, the selection is accepted. And input data associated with all items subordinate to the item are read from the data storage means, and the input data menu display means for displaying the read input data, and the input read by the input data menu display means Until the data is displayed, the selection of one option by the input reception unit, the determination by the determination unit, the generation of the search menu by the item menu generation unit, and the search menu by the input data menu display unit You may comprise so that the input reception means, the determination means, the item menu generation means, and the repetition control means for controlling the input data menu display means may be provided so as to repeat the display.
[0021]
Here, the predetermined criterion in the determination means is that the number of input data associated with the item corresponding to the option for which the selection has been accepted and all items subordinate to the item is equal to or less than a predetermined value. If so, the input data may be displayed, and if it is larger than a predetermined value, it may be configured to be a criterion for generating a search menu.
[0022]
Here, the data storage means stores a data identifier for identifying each input data, and the associating means includes an item reading unit that sequentially reads items included in each classification pattern from the classification pattern storage means. , Sequentially reading input data from the data storage means, and extracting a feature point included in the input data for each read input data; the read item and the extracted feature point When the match determination unit determines whether or not the corresponding item is extracted and the extracted feature points correspond to each other, the read item is determined. In addition, a data association unit for adding a data identifier for identifying the input data including the feature point may be provided.
[0027]
  Here, the classification pattern selection means is input data associated with each item for each classification pattern.A variance calculation unit for calculating a variance value of the input data from the counted value ofSmallest variancevalueAnd a distribution selection unit that selects the calculated classification pattern.
  Here, the classification pattern selection means includes an input data number counting unit that counts the number of types of input data associated with items for each classification pattern, and is associated with each item for each classification pattern. A classification pattern is selected based on a variance calculation unit that calculates a variance value of the input data from a count value of the number of input data, and the number of types of the input data counted and the variance value of the calculated input data And an input data number selection unit.
[0029]
Here, the classification pattern selection unit includes a priority storage unit that stores a priority for each classification pattern, and reads out the priority from the priority storage unit, and uses the read priority to perform one classification. You may comprise so that the priority selection means to select a pattern may be included.
[0030]
Here, the search menu creating device further includes a switch receiving unit that receives a switching instruction for the search menu displayed on the display unit from the user, and the switching receiving unit that receives the switching instruction for the user. A second classification pattern selection means for selecting a classification pattern based on the predetermined criteria in the classification pattern selection means from other classification patterns excluding the classification pattern selected by the classification pattern selection means; The extraction unit may be configured to generate a search menu from the classification pattern selected by the second classification pattern selection unit.
[0031]
Here, the search menu creation device further includes an input reception unit that receives selection of an option from the displayed search menu from the user, and the association unit further corresponds to the option for which the selection has been received. Selection input data extracting means for extracting input data associated with an item to be selected as selection data, correspondence canceling means for canceling all the correspondence between the input data and the item, and an option for accepting the selection A selection input data associating means for associating each item constituting another classification pattern excluding a classification pattern including an item corresponding to the item and the extracted selection data, wherein the classification pattern selection means performs the selection. Based on the predetermined criteria from other classification patterns excluding the classification pattern including the item corresponding to the accepted option. There are, select a classification pattern, the menu extraction means, based on the classification pattern, generates a search menu, the display unit may be configured to display a search menu that is generated.
[0032]
Here, a classification pattern input accepting unit that accepts input of a classification pattern and a classification pattern editing instruction including registration of the classification pattern and deletion of the classification pattern from the user, and an input by the classification pattern input acceptance unit A classification pattern editing unit that edits a classification pattern stored in the classification pattern storage unit based on the classification pattern and the classification pattern editing instruction may be included.
[0033]
DETAILED DESCRIPTION OF THE INVENTION
1. Embodiment 1
Hereinafter, a search menu creating apparatus 100 according to an embodiment of the present invention will be described with reference to the drawings.
1.1 Configuration of search menu creation device
A block diagram showing the configuration of the search menu creation device 100 is shown in FIG.
The search menu creation apparatus 100 includes a data storage unit 101, a data feature point extraction unit 102, a classification pattern storage unit 103, a priority storage unit 104, an association unit 105, a first classification pattern selection unit 106, a menu creation unit 107, a display Unit 108, input unit 109, and second classification pattern selection unit 110.
[0034]
1.1.1 Data storage unit 101
The data storage unit 101 stores input data, for example, message board information 260 sent via a network or the like. In addition, the data storage unit 101 stores feature point information 270 including feature points output from the data feature point extraction unit 102. Specifically, the data storage unit 101 is configured by a magnetic disk device or the like.
A receiving unit (not shown in FIG. 1) receives the input data from the outside via a network or the like, and a data writing unit (not shown in FIG. 1) stores the received input data in the data storage unit 101. Write.
The receiving unit may read input data from a dictionary, thesaurus, or encyclopedia recorded on a computer-readable recording medium such as a CD-ROM.
[0035]
(1) Message board information 260
The message board information 260 is used to convey information that the user wants to buy or sell the article to a large number of people. As shown in FIG. 2, the message board identification information 261, the purchase transfer distinction 262, the address 263, the name of the information provider H.264, age 265, occupation 266, and comment 267.
The message board identification information 261 is a data identifier for identifying the message board information 260 and may be attached in any manner as long as the message board information 260 can be identified. In general, a number determined by the arrival order of input data is used. The purchase transfer distinction 262 indicates whether the information described on the message board is information related to purchase or information related to transfer. The address 263 indicates the address of the information provider, the information provider name 264 indicates the name of the information provider, the age 265 indicates the age of the information provider, and the occupation 266 indicates the occupation of the information provider. . The comment 267 is a sentence including the item name, amount, conditions, etc. relating to purchase or transfer.
[0036]
An example of the message board information 260 stored in the data storage unit 101 is shown in FIG. As shown in this figure, the data storage unit 101 stores message board information 201, 202, and 203 related to purchase and transfer of articles acquired via a network. The message board information 201 includes message board identification information 211, purchase transfer distinction 231, address 232, information provider name 233, age 234, occupation 235, and comment 236. Includes message board identification information 212, purchase / transfer distinction 241, address 242, information provider name 243, age 244, occupation 245, and comment 246. Message board information 203 includes message board identification information. Information 213, purchase / transfer distinction 251, address 252, information provider name 253, age 254, occupation 255, and comment 256 are included.
[0037]
(2) Feature point information 270
The feature point information 270 is composed of feature point identification information 271 and feature points 272 as shown in FIG.
One feature point information 270 corresponds to one message board information 260. The feature point identification information 271 stores the same content as the message board identification information 261 included in the corresponding message board information 260 and is used to identify the feature point information 270. The feature point 272 is extracted by the data feature point extraction unit 102 and written to the data storage unit 101. The feature point 272 will be described later.
An example of the feature point information 270 stored in the data storage unit 101 is shown in FIG. As shown in this figure, the data storage unit 101 stores feature point information 301, feature point information 302, and feature point information 303. The feature point information 301 includes feature point identification information 214 and feature points 221, 222, 223, 224, 225, 226, 227, and 228. The feature point identification information 214 has the same value as the message board identification information 211. , The feature point information 301 corresponds to the message board information 201. The feature point information 302 includes feature point identification information 215 and feature points 281, 282, 283, 284, 285, 286, 287, 288, and 289. The feature point identification information 215 is the same as the message board identification information 212. And the feature point information 302 corresponds to the message board information 202. Further, the feature point information 303 includes feature point identification information 216 and feature points 291, 292, 293, 294, 295, 296, and 297. The feature point identification information 216 has the same value as the message board identification information 213. , The feature point information 303 corresponds to the message board information 203.
[0038]
1.1.2 Data feature point extraction unit 102
For each message board information 260 stored in the data storage unit 101, the data feature point extraction unit 102 includes message board identification information 261 that constitutes the message board information 260, purchase transfer distinction 262, address 263, information provider name 264, and age 265. , Occupation 266 and comment 267 are read out.
Next, the data feature point extraction unit 102 performs morpheme analysis on the read comment 267 to decompose the comment 267 into morphemes, and extracts noun words from the morphemes.
Here, the morpheme analysis is aimed at decomposing a given sentence into a sequence of morphemes which are the smallest linguistic units, and determining the part of speech of each morpheme. ~ 137 (Iwanami Shoten, issued in April 1996), the details are omitted.
[0039]
Further, the data feature point extraction unit 102 uses the read message board identification information 261 as the feature point identification information, the read purchase transfer distinction 262, the address 263, the information provider name 264, the age 265, The occupation 266 and the extracted noun word are written in the data storage unit 101 as feature points.
Here, the purchase transfer distinction 262 included in the message board information 260 indicates whether the information described in the message board is information related to purchase or information related to transfer, and does not include other information. Therefore, as described above, the data feature point extraction unit 102 uses the purchase transfer distinction 262 as it is as a feature point. Similarly, the address 263, the information provider name 264, the age 265, and the occupation 266 are also used as characteristic points. On the other hand, since the comment 267 is a sentence including the name, price, condition, etc. of the item related to purchase or transfer, it is inappropriate to use the comment 267 as a feature point as it is, and as described above, the data feature point The extraction unit 102 performs morphological analysis on the comment 267, extracts a noun word from the comment 267, and uses the extracted noun word as a feature point.
[0040]
In this way, whether the input data is used as a feature point as it is or whether a feature point is extracted by performing morphological analysis on the input data can be determined using the data structure.
[0041]
Next, the data feature point extraction unit 102 will be described using the examples shown in FIGS. 3 and 5.
From the message board information 201 stored in the data storage unit 101, the data feature point extraction unit 102 receives the message board identification information 211 “(1)”, the purchase transfer distinction 231 “sell”, the address 232 “Tokyo”, and the information provider name 233. “Sato”, age 234 “22 years old”, occupation 235 “university student”, comment 236 “PC is sold for 220,000 yen. Next, the data feature point extraction unit 102 performs a morphological analysis on the read comment 236 “PC is sold for 220,000 yen. I would like handing it over”, and the comment 236 is converted into the morpheme “PC”, “ ”,“ 220,000 yen ”,“ de ”,“ sell ”,“ mas ”,“. ”,“ Handing ”,“ ”,“ hope ”,“ mas ”,“. ” Morpheme “PC”, “O”, “220,000 Yen”, “De”, “Sell”, “Masu”, “.”, “Handing”, “O”, “Wish”, “Masu”, The noun words “PC”, “220,000 yen”, and “handing” are extracted from “.”. Further, the data feature point extraction unit 102 uses the read message board identification information 211 “(1)” as the feature point identification information, and reads the purchase transfer distinction 231 “sell”, the address 232 “Tokyo”, and the information provider name 233. “Sato”, age 234 “22 years old”, occupation 235 “university student”, and extracted noun words “computer”, “220,000 yen”, and “handing” are written in the data storage unit 101 as feature points.
The data feature point extraction unit 102 also writes the feature point identification information and the feature point to the data storage unit 101 for the message board information 202 and 203 stored in the data storage unit 101 in the same manner as described above.
[0042]
1.1.3 Classification pattern storage unit 103
The classification pattern storage unit 103 stores classification pattern information 700. Specifically, the classification pattern storage unit 103 is configured by a magnetic disk device or the like.
The classification pattern information 700 is information that serves as a basis for option generation when a search menu including options for selecting input data is generated. The classification pattern information 700 includes at least one classification pattern 701 as shown in FIG.
[0043]
The classification pattern 701 always includes a classification pattern name 702 and an item group 705. The classification pattern 701 may further include a data corresponding item number 703 or a selection mark 704.
The item group 705 includes at least one item 711. The item 711 may be added with at least one message board identification information 713. In addition, the item 711 may be added as a subordinate to at least one small item 712. At least one message board identification information 714 may be added to the small item 712.
The classification pattern name 702 is a name representing the classification pattern 701.
The classification pattern storage unit 103 stores the number of data corresponding items written by the first classification pattern selection unit 106 as the number of data corresponding items 703 for each classification pattern.
The classification pattern storage unit 103 stores a selection mark written by the first classification pattern selection unit 106 as a selection mark 704.
Details of the data correspondence item number 703 and the selection mark 704 will be described later.
[0044]
The message board identification information 713 and 714 is the same as the message board identification information 261 shown in FIG.
Here, any item constituting any one classification pattern is not related to any item constituting any other classification pattern.
FIG. 7 shows classification pattern information 400 as an example of classification pattern information 700 stored in the classification pattern storage unit 103.
The classification pattern information 400 includes classification patterns 421, 431, 441, 451, 461, and 471.
[0045]
The classification patterns 421, 431, 441, 451, 461, and 471 are “age” 411, “occupation” 412, “gender” 413, “place name” 414, “price 1” 415, and “price 2”, respectively. 416 is given.
The classification pattern 421 includes seven items “˜9”, “10-20”, “21-29”, “30-39”, “40-49”, “50-59”, classified into seven ages. An item group 420 composed of “60-” is included.
The classification pattern 431 includes 10 items “infant”, “elementary school student”, “junior high school student”, “high school student”, “college student”, “housewife”, “company employee”, “professional school student”, “self-employed”, “ An item group 430 composed of “unemployed” is included.
The classification pattern 441 includes an item group 440 including two items “male” and “female”.
[0046]
The classification pattern 451 includes an item group 450 including eight regional names “Hokkaido”, “Tohoku”, “Kanto”, “Chubu”, “Kinki”, “China”, “Shikoku”, and “Kyushu”. In addition, the item “Tohoku” 452 includes sub-items “Aomori” 453, “Iwate” 454, “Miyagi” 455, “Akita” 456, “Yamagata” 457, and “Fukushima” 458. Thus, the item has a two-layer structure in which the prefecture name of each region is further set as a small item.
The classification pattern 461 to which the classification pattern name “price 1” 415 is added includes the item group 460 “˜10,000, 1 to 100,000, 100,000 to 1,000,000,” and the classification pattern name “price 2” 416. The classification pattern 471 to which “” is added includes an item group 470 “˜1,000, 1 to 3,000, 3 to 5,000, 5 to 10,000, and 10,000”. As described above, the classification pattern 461 and the classification pattern 471 are different in the configuration of items, although the classification pattern targets are both prices. As described above, a plurality of patterns having the same target but different configurations may be stored.
[0047]
1.1.4 Priority Storage Unit 104
The priority storage unit 104 stores priority information 750 in which priority is associated with each classification pattern 701. Specifically, the priority storage unit 104 is configured by a magnetic disk device or the like.
The priority information 750 includes at least one priority pattern information 751 as shown in FIG. The priority pattern information 751 includes a classification pattern name 752 and a priority 753. The classification pattern name 752 is the same as the classification pattern name 702 shown in FIG. 6, and the priority 753 expresses the degree of preference of the user with respect to the classification pattern indicated by the classification pattern name 752. Indicates that the taste of is high.
[0048]
FIG. 9 shows priority information 760 as an example of the priority information 750. As shown in this figure, the classification pattern names “age” 511, “occupation” 512, “gender” 513, “place name” 514, “price 1” 515, “price 2” 516 have priority “3” 521, respectively. , “4” 522, “5” 523, “8” 524, “2” 525, and “4” 526 are associated with each other. Accordingly, the priority levels “3”, “4”, “5”, “8”, “2”, and “4” are assigned to the classification patterns 421, 431, 441, 451, 461, and 471 shown in FIG. It will be received.
In this embodiment, since the classification pattern name is associated with the priority, it is possible to know the priority of the classification pattern using the classification pattern name, but the classification pattern name 752 is used instead of the classification pattern name 752 of the priority information 750. The storage position of the pattern may be stored. In this case, the classification pattern name 700 is not provided in the classification pattern information 700 but the storage position of the classification pattern is provided.
Further, the priority may be set higher as the priority number is smaller.
[0049]
1.1.5 Corresponding part 105
The association unit 105 uses the feature point information stored in the data storage unit 101 to associate the message board information stored in the data storage unit 101 with the classification pattern stored in the classification pattern storage unit 103.
A specific operation of the associating unit 105 will be described with reference to FIGS. 5 and 7.
The association unit 105 first reads the feature point identification information 214 and the feature point 221 included in the feature point information 301 from the data storage unit 101, and the read feature point 221 is stored in the classification pattern storage unit 103. Compare with items or sub-items included in the patterns 421, 431, 441, 451, 461, 471. If the feature point 221 matches the item or the small item, or if the feature point 221 is included in the range indicated by the item or the small item, the associating unit 105 determines whether the feature point 221 matches or The read feature point identification information 214 is added as message board identification information to a small item or an item or small item in which the feature point 221 is included in the range. Thereby, the feature point information is associated with the item or the small item.
[0050]
Corresponding unit 105 compares feature points with classification pattern items or small items for feature points 222, 223, 224, 225, 226, 227, and 228 included in feature point information 301, as with feature points 221. And adding the feature point identification information 214.
[0051]
Next, the associating unit 105 compares the feature points included in the feature point information 302 with the feature points and classification pattern items or small items, and the feature point identification information 215, as with the feature point information 301. Is added. Similarly, the associating unit 105 compares the feature points with classification pattern items or small items and adds feature point identification information for all feature point information included in the data storage unit 101.
For example, the associating unit 105 reads the feature point identification information 214 “(1)” and the feature point 221 “sell” included in the feature point information 301, and the feature points 221 and the classification patterns 421, 431, 441, 451. , 461, 471 and the items included in the sub items. The items or small items included in the classification patterns 421, 431, 441, 451, 461, 471 have no item or small item that matches the feature point 221 “sell”, and the feature point 221 “sell” Since it is not included in the range indicated by the item or the small item, the associating unit 105 reads the next feature point 222 “Tokyo”.
[0052]
The associating unit 105 compares the feature point 222 “Tokyo” with the items or small items included in the classification patterns 421, 431, 441, 451, 461, and 471. Since the feature point 222 “Tokyo” matches the small item “Tokyo” included in the classification pattern 451, the feature point identification information 214 “(1)” is added to the small item “Tokyo” included in the classification pattern 451. Add as information.
FIG. 10 shows classification pattern information when the feature point identification information 214 “(1)” is added as message board identification information to the small item “Tokyo” included in the classification pattern 451. As shown in this figure, the code | symbol 611 has shown this matched message board identification number.
Similarly, for the other feature points 223, 224, 225, 226, 227, and 228 included in the feature point information 301, the feature point identification information is added to the item or sub-item.
Here, since the feature point 223 “Sato”, the feature point 226 “PC”, and the feature point 228 “handing” are not in the classification pattern, the feature point identification information is not added to the item or the small item.
[0053]
Further, numerical data such as the feature point 227 “220,000” is determined by using the magnitude relation of the numerical values to determine whether they are within the range of the item. In the present embodiment, message board information is included in items included in both of the classification patterns 461 and 471 indicated by the two classification pattern names 415 “price 1” and 416 “price 2” for the feature point 227 “220,000”. 201 are associated with each other.
FIG. 11 shows the state of the classification pattern storage unit 103 when the association using all the feature points included in the feature point information 301 is completed. In this figure, reference numerals 611 to 615 indicate message board identification information 211 “(1)” of the message board information 201 associated by the associating unit 105.
FIG. 12 shows the classification pattern information 652 when the association using all the feature points included in the feature point information 301, 302, 303 is completed. Reference numerals 811 to 822 indicate message board identification numbers of the message board information 201, 202, and 203 associated with each other.
In FIG. 13, another 17 feature point information is newly added to the feature point information 301, 302, 303, and the association using all the feature points included in the total 20 feature point information is completed. The classification pattern information 653 is shown.
[0054]
1.1.6 First classification pattern selection unit 106
The first classification pattern selection unit 106 uses priority information stored in the priority storage unit 104 from classification pattern information stored in the classification pattern storage unit 103 based on a predetermined criterion described below. Select the optimal classification pattern for creating a search menu.
The first classification pattern selection unit 106 reads out the classification pattern information 700 stored in the classification pattern storage unit 103 for each classification pattern 701, and items and message boards to which the message board identification information 713 is added for each of the read classification patterns 701. The number of items including small items to which the identification information 714 is added is counted, and the counted number of items is written in the classification pattern information 700 of the classification pattern storage unit 103 for each classification pattern 701 as the data-corresponding item number 703.
[0055]
Here, the operation of the first classification pattern selection unit 106 will be described using the classification pattern information 653 shown in FIG.
The first classification pattern selection unit 106 reads the classification pattern 1010 from the classification pattern information 653, determines whether or not message board identification information is added to each item included in the classification pattern 1010, and the message board identification information is added. Count the number of items you have.
Among the items included in the classification pattern 1010, the input data is associated with two items 1012, 1013. Therefore, the first classification pattern selection unit 106 counts “2” as the number of items, and the data “2” is written in the classification pattern 1010 as the number of corresponding items.
Here, for a classification pattern having a two-level structure of items and small items indicated by the classification pattern name “place name” 1021, the first classification pattern selection unit 106 adds the message board identification information to the small items. Count the number of items that contain.
FIG. 14 shows a state in which the number of items associated with message board identification information is stored for each classification pattern name.
[0056]
When creating a search menu, if the number of options displayed per screen is too small, the menu hierarchy becomes deep and difficult to search, and conversely if too many, it is difficult to compare at a time, making selection difficult. Therefore, a predetermined number of the best options is set in advance, and the closer to the number, the more appropriate classification is determined. Here, assuming that the number of appropriate options is 6, in the example shown in FIG. 14, the number of items of the classification pattern indicated by the classification pattern name 1021 “place name” and 1022 “price 2” is 5, respectively. Therefore, the first classification pattern selection unit 106 determines that the classification pattern indicated by the classification pattern name 1021 “place name” and 1022 “price 2” is optimal.
[0057]
In this way, the first classification pattern selection unit 106 selects an optimal classification pattern. Here, when there are a plurality of optimum classification patterns, as shown below, the first classification pattern selection unit 106 further narrows down the selected plurality of classification patterns and selects a classification pattern.
That is, the first classification pattern selection unit 106 reads a plurality of priorities corresponding to the classification pattern names of the plurality of classification patterns selected as the optimum from the priority storage unit 104, and sets the plurality of read priorities. Compare. For example, as illustrated in FIG. 9, the first classification pattern selection unit 106 has a priority 524 “8” corresponding to the classification pattern name 1021 “place name” and a priority 526 corresponding to the classification pattern name 1022 “price 2”. Read “4”. The first classification pattern selection unit 106 compares the read priority 524 “8” with the priority 526 “4”. Since the priority 524 “8” is higher than the priority 526 “4”, the classification pattern name 1021 The classification pattern indicated by “place name” is determined to be optimal.
When the priorities are also equal, the first classification pattern selection unit 106 selects one classification pattern using a method such as determining at random or determining in the order of registration of the classification patterns.
[0058]
The first classification pattern selection unit 106 adds a selection mark “◯” indicating that the classification is determined to be optimal to the classification pattern 701 in the classification pattern storage unit 103 as a selection mark 704 and writes the selection pattern “◯”. .
FIG. 15 shows classification pattern information 1100 when 1111 “◯” is added as a selection mark to the determined classification pattern name “place name”.
[0059]
1.1.7 Menu creation unit 107
The menu creation unit 107 extracts only items associated with message board information from the classification pattern selected by the first classification pattern selection unit, and creates a search menu using the extracted items as options. Details of the operation of the menu creation unit 107 will be described below with reference to the drawings.
The menu creation unit 107 extracts the classification pattern 701 including the selection mark 704 of “O” from the classification pattern information 700 stored in the classification pattern storage unit 103, and the message board identification information 713 from the extracted classification pattern 701. And items including small items to which message board identification information 714 is added are extracted. In addition, the menu creation unit 107 calculates the number of message board identification information 713 added to the item for each of the extracted items, and calculates the number of message board identification information 714 added to the small item included in the item. The calculated number of message board identification information 713 and the calculated number of message board identification information 714 are totaled, and the sum is obtained as the number of message board information added to the item.
In this way, the menu creation unit 107 generates at least one set of the extracted item and the number of message board information added to the item, and the generated at least one set of the item and the message board information Are output to the display unit 108.
[0060]
Here, the details of the operation of the menu creation unit 107 will be specifically described with reference to FIG.
First, the menu creation unit 107 extracts a classification pattern 1120 with a selection mark “◯” 1111. Next, the menu creation unit 107 includes items 1121 “Hokkaido”, 1122 “Tohoku”, 1123 “Kanto”, 1124 “Chubu”, 1125 “Kinki”, 1126 “China”, 1127 “Shikoku” included in the classification pattern 1120. Items 1123 “Kanto”, 1124 “Chubu”, 1125 “Kinki”, 1126 “China”, 1127 “Shikoku” including small items associated with message board identification information from the respective items Is extracted. In this example, there is no item to which message board identification information is added. The menu creation unit 107 identifies the message board added to the small items included in the five items 1123 “Kanto”, 1124 “Chubu”, 1125 “Kinki”, 1126 “China”, and 1127 “Shikoku” extracted in this way. The number of information is calculated for each item. In this way, the menu creation unit 107 generates five sets of the extracted item and the number of message board information added to the item. FIG. 16 shows five sets of the item 1201 generated by the menu creation unit 107 and the number 1202 of message board information. Next, the menu creation unit 107 outputs five sets of the extracted items and the number of message board information calculated for each item to the display unit 108.
[0061]
1.1.8 Display unit 108
The display unit 108 receives at least one set of items and the number of message board information from the menu creation unit 107, and based on at least one set of the received items and the number of message board information, the received items and At least one set of a bar graph having a length proportional to the number of received message board information and the number of received message board information is displayed as a search menu. The display unit 108 displays a cursor for selecting an item.
FIG. 17 shows an example of the search menu displayed by the display unit 108. In this figure, the search menu is composed of five sets of an item 1302, a bar graph 1303, and a number 1304 of message board information, and a cursor 1301 for selecting an item is displayed.
In addition, although the display part 108 is displaying the item, a bar graph, and the number of message board information, you may display an item and the number of message board information.
[0062]
When the display unit 108 receives “move up” from the input unit 109 in a state where the search menu is displayed on the display unit 108, the display unit 108 moves the cursor 1301 to the next item and displays the item. When “downward movement” is received, the cursor 1301 is moved to the next lower item and displayed.
When “execute” is received from the input unit 109, the display unit 108 selects the item indicated by the cursor 1301, and stores one message board identification information added to the selected item in the classification pattern storage unit 103. The message board information 260 identified by the one message board identification information read out from the classified pattern information 700 is read out from the data storage unit 101, and the read out message board information 260 is displayed.
When the display unit 108 receives the “downward movement” from the input unit 109 while the display unit 108 displays the message board information 260, the read message board identification information added to the selected item follows. Further, another message board identification information is read from the classification pattern information 700 stored in the classification pattern storage unit 103, and the message board information 260 identified by one read message board identification information is read from the data storage unit 101 and read. Message board information 260 is displayed.
[0063]
When the display unit 108 displays “message board information 260” and the display unit 108 receives “move up” from the input unit 109, the message board identification information before the read message board identification information added to the selected item is displayed. Further, another message board identification information is read from the classification pattern information 700 stored in the classification pattern storage unit 103, and the message board information 260 identified by one read message board identification information is read from the data storage unit 101 and read. Message board information 260 is displayed.
When the display unit 108 receives “execute” from the input unit 109 while the search menu shown in FIG. 17 is being displayed, the display unit 108 adds the item “Kanto” included in the classification pattern information. The read message board identification information “(1)” is read from the classification pattern storage unit 103, the message board information 201 identified by the read message board identification information “(1)” is read from the data storage unit 101, and the read message board is read. Information 201 is displayed. FIG. 19 shows a state in which the message board information 201 is displayed on the display unit 108. When the user presses the up cursor key 1401 and the down cursor key 1402 in this state, other message board information is read and displayed.
[0064]
1.1.9 Input unit 109
As shown in FIG. 18, the input unit 109 includes an up cursor key 1401, a down cursor key 1402, an execute key 1403, and a switch key 1404, and accepts user input.
When the up cursor key 1401 is pressed by the user while the search menu is displayed on the display unit 108, the input unit 109 outputs “move up” to the display unit 108. When the down cursor key 1402 is pressed by the user while the search menu is displayed on the display unit 108, the input unit 109 outputs “move down” to the display unit 108.
When the execution key 1403 is pressed by the user, the input unit 109 outputs “execute” to the display unit 108.
When the switch key 1404 is pressed by the user, the input unit 109 outputs “switch” to the second classification pattern selection unit 110.
The input unit 109 includes four up cursor keys 1401, a down cursor key 1402, an execute key 1403, and a switch key 1404. The functions of the four keys can be assigned to any key of a general computer keyboard. You may make it allocate.
[0065]
1.1.10 Second classification pattern selection unit 110
When the display unit 108 displays a search menu as shown in FIG. 17, the second classification pattern selection unit 110 is used to create another search menu when the user does not like the classification based on the search menu. It is done.
When receiving the “switch” from the input unit 109, the second classification pattern selection unit 110 detects the selection mark 704 of “◯” from the classification pattern information 700 stored in the classification pattern storage unit 103, and the detected selection mark The “◯” in 704 is changed to “●” indicating used, and is written in the classification pattern information 700.
In the classification pattern information 1100 shown in FIG. 15, “◯” 1111 is added to the selection mark of the currently selected classification pattern 1120. The second classification pattern selection unit 110 reads “◯” 1111 of this selection mark, changes the read selection mark to “●” indicating that it has been used, and writes it again into the classification pattern information 1100. FIG. 20 shows classification pattern information 1600 when “●” 1601 indicating used is attached to the selection mark.
[0066]
Next, the second classification pattern selection unit 110 selects an optimum classification pattern by the same method as the first classification pattern selection unit 106 from the classification patterns not marked with “●” indicating used in the selection mark. Taking FIG. 21 as an example, the second classification pattern selection unit 110 selects the classification pattern 1612 indicated by the classification pattern name 1611 “Price 2”.
The second classification pattern selection unit 110 puts “◯” on a selection mark indicating that the selected classification pattern is selected. FIG. 21 shows classification pattern information 1620 when a selection mark indicating that it has been selected is marked with “◯”. “◯” 1602 is attached to the classification pattern 1612 as a selection mark indicating that it has been selected.
[0067]
The second classification pattern selection unit 110, if there is a classification pattern that can be selected, that is, a classification pattern without a used mark “●”, every time the switch key 1404 is pressed many times, A classification pattern can be selected. In the case of the classification pattern information 1620 shown in FIG. 21, it can be switched three more times.
In this way, when the second classification pattern selection unit 110 selects a classification pattern, the menu creation unit 107 again uses the classification pattern information including the classification pattern with “◯” as a selection mark to select a search menu. The display unit 108 displays the search menu.
[0068]
1.2 Operation of search menu creation device 100
Here, the operation of the search menu creation device 100 will be described.
1.2.1 Search menu creation operation
Here, the operation in which the search menu creating apparatus 100 creates the search menu will be described with reference to the flowchart shown in FIG.
[0069]
The data feature point extraction unit 102 extracts feature points from the message board information 260 stored in the data storage unit 101, and writes the extracted feature points into the data storage unit 101 as feature point information 270 (step S1801). The associating unit 105 extracts the feature point identification information 271 from the feature point information 270 stored in the data storage unit 101 (step S1802), and extracts one feature point 272 included in the feature point information 270 (step S1803). Then, the extracted one feature point 272 is compared with the items included in the classification pattern information stored in the classification pattern storage unit 103 (step S1804), and if one feature point 272 matches the item, The extracted feature point identification information 271 is added to the item and associated (step S1805). Extraction, comparison, and association of feature points are repeated for all feature points included in the feature point information 270 (steps S1804 to S1806). Further, extraction of feature point identification information, extraction of feature points, comparison, and association are repeated for all feature point information included in the data storage unit 101 (steps S1802 to S1807). When the association is completed for all feature point information, the first classification pattern selection unit 106 calculates the number of items associated with the message board identification information for each classification pattern (step S1808), and sets it in advance. A classification pattern having the number of items close to the optimal number of items is selected as the optimal classification pattern (step S1809). Here, when there are a plurality of optimum classification patterns, the first classification pattern selection unit 106 further refers to the priority storage unit 104 and selects one of the selected classification patterns with the highest priority. A selection mark “◯” indicating selection is added to the selected classification pattern (step S1810). Next, the menu creation unit 107 creates a search menu using items included in the selected classification pattern as options (step S1811), and the display unit 108 displays the created search menu (step S1812).
[0070]
1.2.2 Operation when the switch key is pressed
Here, the operation in the case where the search key creation device 100 presses the switch key to create the search menu again will be described with reference to the flowchart shown in FIG.
In the flowchart shown in FIG. 23, steps S1811 and S1812 are the same as the steps denoted by the same reference numerals in the flowchart shown in FIG.
[0071]
When the switch key 1404 is pressed by the user, the input unit 109 outputs “switch” to the second classification pattern selection unit 110 to instruct switching of the search menu screen (step S1901), and the second classification pattern selection unit. 110 adds a selection mark “●” indicating used to the currently selected classification pattern (step S1902). Next, the second classification pattern selection unit 1903 selects the classification pattern having the number of items close to the preset optimum number of items from the unused classification patterns without the selection mark “●” indicating used. Is selected as the optimum classification pattern (step S1903). Here, when there are a plurality of optimum classification patterns, the second classification pattern selection unit 110 further refers to the priority storage unit 104 and selects one of the selected classification patterns with the highest priority. A selection mark “◯” indicating selection is added to the selected classification pattern (step S1904). Subsequently, the menu creation unit 107 creates a search menu using the classification pattern to which the selection mark “◯” is added, and the display unit 108 displays the search menu (steps S1811 to S1812).
[0072]
2. Embodiment 2
Hereinafter, a search menu creating apparatus 2100 according to another embodiment of the present invention will be described with reference to the drawings.
2.1 Configuration of Search Menu Creation Device 2100
FIG. 24 is a block diagram showing the configuration of the search menu creation apparatus 2100.
The search menu creation device 2100 includes a data storage unit 2101, an important word extraction unit 2102, an important word storage unit 2103, a classification pattern storage unit 2104, an association unit 2105, a classification pattern selection unit 2106, a menu creation unit 2107, a display unit 2108, An input unit 2109 is included.
[0073]
2.1.1 Data storage unit 2101
The data storage unit 2101 is configured by a magnetic disk device or the like. The data storage unit 2101 adds an identification number to document data sent via a network or the like, and stores the document data to which the identification number is added. The identification number is a number for specifying the document data. The identification numbers are assigned in the order in which the document data is received.
As shown in FIG. 25, the data storage unit 2101 stores document data 2201, 2202,. Document data 2201, 2202,... Are newspaper articles obtained through a network. Further, identification numbers “1” 2211, “2” 2212,... For identifying the document data are added to the document data 2201, 2202,.
[0074]
2.1.2 Important word extraction unit 2102
The keyword extraction unit 2102 reads the document data stored in the data storage unit 2101, performs morphological analysis on the read document data, and decomposes the document data into at least one morpheme.
When the document data is written in Japanese, the keyword extraction unit 2102 extracts at least one independent word (in English, corresponding to a full word or a content word) from the decomposed at least one morpheme. To do. If the document data is written in English, the keyword extraction unit 2102 has an unnecessary word list (stop words in English) and registers words that are not to be an important word in the unnecessary word list. . Specifically, prepositions, be verbs, and the like are registered. Next, morphemes registered in the unnecessary word list are excluded from the decomposed morphemes.
[0075]
Next, the important word extraction unit 2102 counts the number of times the same independent word appears in the extracted at least one independent word, and a predetermined number of independent words in order from the independent word from which the largest number of independent words has been counted. And a predetermined number of extracted independent words are set as important words as an example of feature points.
The important word extraction unit 2102 reads the identification number added to the document data stored in the data storage unit 2101.
Further, the keyword extraction unit 2102 outputs the read identification number and a predetermined number of independent words as the keyword to the keyword storage unit 2103.
[0076]
The keyword extraction unit 2102 extracts the keyword and reads the identification number for all the document data stored in the data storage unit 2101, as described above, The predetermined number of independent words is output to the important word storage unit 2103.
Independent words here are words that convey a concept or an image that have a substantial meaning, such as most nouns, adjectives, verbs, and adverbs.
The operation of the keyword extraction unit 2102 will be described below using the document data 2201 shown in FIG. 25 as an example.
[0077]
The keyword extraction unit 2102 performs morphological analysis on the document data 2201 stored in the data storage unit 2101 and decomposes the document data 2201 to obtain a plurality of words as shown in FIG. In the figure, “/” represents a word break. Next, the important word extraction unit 2102 extracts a plurality of independent words as important word candidates from the plurality of words shown in FIG. 26, and calculates the number of times the same independent word appears in the extracted plurality of independent words. Count. FIG. 27 shows 10 words that are extracted independent words and the number of appearances of those words. Here, the important word extraction unit 2102 extracts only five independent words having a high frequency of appearance from ten words that are independent words extracted as important word candidates, and sets them as important words. In this way, five words “sunny”, “weather”, “Osaka”, “Kyoto”, and “Kobe” are extracted as important words.
In this embodiment, the keyword extraction unit 2102 extracts the keyword using only the number of occurrences of the word in the document data. However, the present invention is not limited to this.
[0078]
The keyword extraction unit 2102 may determine the keyword based on the appearance position of the word. For example, a word appearing at the beginning of a title or text in document data may be regarded as an important word.
Further, the important word extraction unit 2102 may determine an important word according to the role of the word. For example, morphological analysis is performed on document data, morphemes are extracted, syntax analysis is performed on the extracted morphemes to clarify the sentence structure of the document data, and a certain word is the subject in the sentence structure. If the word is a subject, it may be regarded as an important word.
[0079]
Further, the important word extraction unit 2102 compares the number of appearances of a certain word in general document data with the number of appearances of the word in the document data to be searched, so that the important word of the word is obtained. It may be. Specifically, the important word extraction unit 2102 regards the word as an important word when a word with a small number of appearances in general document data appears frequently in the document data to be searched. May be.
In this embodiment, the number of important words is 5. However, other numbers such as 4, 3, 6, 7 can be used. Further, instead of a fixed number, all independent words that are equal to or more than a predetermined number of appearances may be extracted, and the extracted independent words may be used as important words.
[0080]
2.1.3 Important word storage unit 2103
The important word storage unit 2103 receives the identification number and a predetermined number of important words from the important word extraction unit 2102 for each document data, and stores the received identification number and the predetermined number of important words as important word information 2530. Since the important word storage unit 2103 receives the same number of identification numbers as the number of document data stored in the data storage unit 2101 and a predetermined number of important words, the important word storage unit 2103 stores the identification number in the data storage unit 2101. The same number of important word information 2530 as the number of document data being stored is stored. The keyword storage unit 2103 is specifically composed of a memory or the like.
FIG. 28 shows the configuration of the keyword information 2530. The important word information 2530 includes important word identification information 2531 and a predetermined number of important words 2532. Important word identification information 2531 is the received identification number, and a predetermined number of important words 2532 are the received predetermined number of important words.
Here, the predetermined number is extracted when the important word extraction unit 2102 counts the number of times the same independent word appears, and extracts a predetermined number of independent words in order from the independent word that has been counted the most. The number of independent words.
[0081]
FIG. 29 shows an example of important word information 2530 stored in the important word storage unit 2103. As shown in this figure, the important word storage unit 2103 stores important word information 2506, 2516,..., And the important word information 2506 includes important word identification information 2521 and important words 2501-2505. The important word information 2516 includes important word identification information 2522 and important words 2511 to 2515.
Here, the important words 2501 to 2505 are important words extracted from the document data 2201, respectively, and the important words 2511 to 2515 are important words extracted from the document data 2202, respectively. The important word identification information 2521 has the same content as the identification number 2211 of the document data 2201, and the important word identification information 2522 has the same content as the identification number 2212 of the document data 2202. Accordingly, the important word information 2506 corresponds to the document data 2201, and the important word information 2516 corresponds to the document data 2202.
[0082]
2.1.4 Classification pattern storage unit 2104
The classification pattern storage unit 2104 stores classification pattern information 2700 for classifying document data with keywords. Specifically, the classification pattern storage unit 2104 is configured by a magnetic disk device or the like.
The classification pattern information 2700 includes at least one classification pattern 2701 as shown in FIG.
The classification pattern 2701 includes a selection mark 2711, a classification pattern number 2712, a classification viewpoint name 2713, a document data correspondence number 2714, and at least one classification item 2715. The classification item 2715 includes a classification item name 2721, a document data number 2722, The document data identification number 2723 and at least one keyword 2724 are included.
[0083]
The selection mark 2711 is the same as the selection mark 704, and a description thereof will be omitted.
The classification pattern number 2712 is a number for identifying the classification pattern 2701.
The classification viewpoint name 2713 is a name representing the classification pattern 2701.
The document data correspondence number 2714 is the number of document data types identified by all the document data identification numbers 2723 included in the classification pattern 2701.
The classification item name 2721 is a name representing the classification item 2715.
The document data number 2722 is the number of document data identification numbers included in the classification item 2715.
The document data identification number 2723 is an identification number for identifying document data stored in the data storage unit 2101.
The keyword 2724 is used for associating the classification item with the document data.
[0084]
FIG. 31 shows classification pattern information 2600 as an example of classification pattern information 2700 stored in the classification pattern storage unit 2104. As shown in this figure, the classification pattern storage unit 2104 stores classification pattern information 2600, and the classification pattern information 2600 includes a classification pattern 2601, a classification pattern 2602, a classification pattern 2603, and so on. ing.
In the classification pattern 2601, “8 regions in Japan” 2611 is a classification viewpoint. “1” 2612 is a classification pattern number for identifying a classification pattern. The classification pattern 2601 includes classification items 2631, 2632, 2633, 2634, 2635, 2636, 2637, and 2638.
In the classification item 2631, “Hokkaido” 2621 is a classification item name, “Hokkaido” 2641, “Sapporo” 2642, “Asahikawa” 2643, “Hakodate” 2644, “Kushiro” 2645, “Otaru” 2646, “Tomakomai” 2647 is a keyword.
Similarly, the classification item 2632 includes a classification item name and a plurality of keywords. The same applies to the classification items 2633, 2634, 2635, 2636, 2637, and 2638.
[0085]
2.1.5 Corresponding part 2105
(1) Correspondence between key words, classification patterns and classification items
The association unit 2105 reads the keyword “sunny” 2501 from the keyword information 2506 stored in the keyword storage unit 2103.
Next, the associating unit 2105 reads the keyword “Hokkaido” 2641 that constitutes the classification item 2631 included in the classification pattern 2601 in the classification pattern information 2600 stored in the classification pattern storage unit 2104, and the read important word “Sunny” 2501 is compared with the keyword “Hokkaido” 2641. Since the keyword “Hokkaido” 2641 and the keyword “sunny” 2501 do not match, the associating unit 2105 next reads the keyword “Sapporo” 2642 that constitutes the classification item 2631, and the keyword “sunny” 2501 and the keyword “ Compare with Sapporo 2642. Similarly, the keywords constituting the classification item 2631 are sequentially read out until a keyword matching the important word “sunny” 2501 is found, and the read keyword and the important word “sunny” 2501 are compared.
[0086]
None of the plurality of keywords constituting the classification item 2631 matches the important word “sunny” 2501. Next, the associating unit 2105 sequentially reads out the keywords and compares them with the keyword “sunny” 2501 until a keyword matching the keyword “sunny” 2501 is found for a plurality of keywords constituting the classification item 2632. I do.
None of the plurality of keywords constituting the classification item 2632 matches the important word “sunny” 2501, and therefore the associating unit 2105 relates to the classification items 2633 to 2638 until a keyword matching the important word “sunny” 2501 is found. For a plurality of keywords constituting each classification item, the keyword reading is sequentially performed and a comparison with the keyword “sunny” 2501 is performed.
[0087]
The associating unit 2105 could not find a keyword that matches the important word “sunny” 2501 in the classification pattern 2601. Next, for the classification patterns 2602, 2603,..., The keyword reading is compared with the keyword “sunny” 2501. As a result, the associating unit 2105 cannot find a keyword that matches the keyword “sunny” 2501 in the classification pattern information 2600, and the associating unit 2105 stores the keyword in the keyword storage unit 2103. The important word “weather” 2502 is read from the important word information 2506, and the keyword matching the read important word “weather” 2502 is searched from the classification pattern information 2600. As a result, the associating unit 2105 includes the keyword “weather” 2502 among the keywords included in the sixth classification item in the classification pattern indicated by the classification pattern number (14) (not shown in FIG. 31). A keyword matching with was detected. Therefore, the associating unit 2105 stores the keyword “weather” 2502, the group of the classification pattern number (14), and the sixth classification item in association with each other.
[0088]
Similarly, the associating unit 2105 sequentially reads the important words “Osaka” 2503, “Kyoto” 2504, and “Kobe” 2505 included in the important word information 2506 stored in the important word storage unit 2103 and reads the important words If keywords matching the words “Osaka” 2503, “Kyoto” 2504, and “Kobe” 2505 are searched from the keywords included in the classification pattern information 2600, and a matching keyword is found, the keyword, classification pattern number, The classification items are stored in association with each other.
In this manner, when the association between all the important words included in one important word information stored in the important word storage unit 2103 and all the keywords included in the classification pattern information 2600 is completed, FIG. As shown, the association unit 2105 stores the classification pattern number (14) 2821 and the classification item (6) 2822 in association with each other for the important word “weather” 2802 and stores the important word “Osaka” 2803. , Classification pattern number (1) 2831, classification item (5) 2832, classification pattern number (12) 2833, and classification item (2) 2834 are stored in association with each other. Number (1) 2841 and classification item (5) 2842 are stored in association with each other, and for important word “Kobe” 2805, classification pattern number (1) 28 1 and classification item (5) 2852 and in association with each other is stored. It should be noted that no association has occurred for the important word “sunny” 2801.
[0089]
(2) Duplicate deletion of pairs
As shown in FIG. 32, the group in which the classification pattern number (1) and the number (5) for identifying the classification item name are associated has appeared three times. As shown in FIG. 33, the association 2861 between the classification pattern number (14) and the number (6) for identifying the classification item name, the classification pattern number (1), and the classification item name are identified. The association 2862 with the number (5) and the association 2863 with the classification pattern number (12) and the number (2) for identifying the classification item name are stored.
[0090]
(3) Association of document identification numbers
The association between the classification pattern number and the number for identifying the classification item name shown in FIG. 33 is performed using the important words extracted from the document data 2201. Therefore, the association unit 2105 As shown in FIG. 34, in order to associate the classification item 2635 in the classification pattern 2601 with the document data and the identification number “(1)” of 2201, the identification number “(1)” 2901 is added to the classification item 2635.
[0091]
(4) Writing the number of document data
The association unit 2105 calculates the number of identification numbers of the associated document data for each classification item, and writes it in the classification item of the classification pattern for each classification item.
Also, the associating unit 2105 calculates the number of document data types associated with the classification pattern for each classification pattern, and calculates the calculated number of document data types for each classification pattern. Is stored in the classification pattern. Here, calculation is performed assuming that the document data associated with each other is duplicated.
FIG. 35 shows the classification pattern 2601 after the association is completed as described above. In this figure, keywords corresponding to classification item names are not shown.
In this figure, reference numerals 3113 to 3118 are identification numbers of the associated document data, reference numerals 3102 to 3109 are the document data numbers associated with the respective classification items, and reference numeral 3101 is the document data correspondence number. is there. Since the document data identification number (3) is associated with the three classification items, the sum of the number of document data associated with the classification item is 14, but if the duplicate document data is excluded, the classification pattern The number of types of document data associated with is 12, and the number of document data correspondence is 12.
[0092]
2.1.6 Classification pattern selection unit 2106
Since the classification pattern selection unit 2106 selects a classification pattern using the association result of the document data stored in the classification pattern information, the classification pattern evaluation unit 2106 evaluates the classification pattern using the evaluation formula (Formula 1) shown in FIG. I do.
In (Expression 1), P is an evaluation value of the classification pattern, and n in the first term of (Expression 1) is the number of document data associated with the classification pattern. The second term m in (Expression 1) is the number of classification item names associated with the document data. The function f (x) is a function having the number of classification items as an argument as shown in FIG. 36 (b), and takes a maximum value of 1 when x = 5 and takes a minimum value of 0 when x = 0. When x = 10, the minimum value is 0. The square of σ in the third term of (Equation 1) is the variance of the number of associated document data. α, β, and γ are constants.
[0093]
The first term of (Expression 1) indicates that the evaluation value of the classification pattern increases as the number of document data associated with the classification pattern increases. The second term indicates that the evaluation value of the classification pattern increases as the number of classification items associated with the document data in the classification pattern approaches five. The third term indicates that the evaluation value of the classification pattern increases as the variance of the number of document data associated with the classification pattern decreases.
An evaluation formula (formula 1) is applied to the classification pattern 2601 shown in FIG. 35 to calculate the evaluation value of the classification pattern. Here, the constants α, β, and γ are each 1. Since the document data correspondence number is 12, the value of the first term of the evaluation formula (Formula 1) is 12.
Among the classification items of the classification pattern 2601, there are six classification items associated with the document data. Therefore, when the function f (x) shown in FIG. The second term is 0.8.
Next, with respect to the third term of the evaluation formula (Formula 1), only the classification items associated with the document data are targeted, and first, (1 + 2 + 3 + 5 + 2 + 1) /6=2.33 as the average of the number of corresponding document data. Get. Next, as the variance of the number of corresponding document data, σ² = (1-2.33)²+ (2-2.33)²+ (3-2.33)²+ (5-2.33)²+ (2-2.33)²+ (1-2.33)²= 11.33 is obtained. Thus, the third term becomes 1 / (1 + 11.33) = 0.08.
Therefore, the evaluation value of the classification pattern 2601 is 12 + 0.8 + 0.08 = 12.88.
Here, for convenience, the values of the constants α, β, and γ are each 1. However, these constants are constants for balancing each term, and appropriate values can be set.
[0094]
As described above, according to the first term of the evaluation formula (Formula 1), it is possible to create a search menu option preferentially from a classification pattern associated with a large amount of input data, and to create an effective search menu. is there. According to the second term of the evaluation formula (Formula 1), menu options can be created preferentially from a classification pattern with an appropriate number of items associated with input data, and an effective search menu can be created. . According to the third term of the evaluation formula (Formula 1), it is possible to preferentially create menu choices from a classification pattern in which the number of input data associated with each item is as uniform as possible, thereby creating an effective search menu. There is.
As described above, the classification pattern selection unit 2106 calculates the evaluation value of the classification pattern for every classification pattern, and selects the classification pattern having the maximum evaluation value as the optimum classification pattern.
The classification pattern selection unit 2106 adds a selection mark “◯” to the classification pattern selected as the optimal classification pattern.
[0095]
2.1.7 Menu creation unit 2107
The menu creation unit 2107 extracts the classification pattern to which the selection mark “◯” has been added by the classification pattern selection unit 2106 from the classification pattern storage unit 2104, and the classification viewpoint name and the document data identification number are extracted from the extracted classification pattern. And extracting at least one set of the classification item name included in the extracted classification item and the number of document data associated with the extracted classification item, and the classification viewpoint name and The extracted at least one set is output to the display unit 2108. If the classification pattern 2601 shown in FIG. 35 is selected, the menu creation unit 2107 creates a set of a classification viewpoint name, a classification item name, and the number of document data, as shown in FIG.
[0096]
2.1.8 Display unit 2108
The display unit 2108 receives the classification viewpoint name, the classification item name, and at least one set of the number of document data from the menu creation unit 2107. The received classification viewpoint name, the received classification item name, a bar graph having a length proportional to the received document data number, and at least one set of the received document data number are displayed as a search menu. The display unit 2108 displays a cursor for selecting a classification item.
FIG. 37B shows an example in which the search menu created by the display unit 2108 is displayed. In this figure, a cursor 3201 for selecting input data is displayed.
[0097]
When the display unit 2108 receives “move up” from the input unit 2109, it moves the cursor 3201 to the next higher category and displays it. When it receives “move down” from the input unit 2109, the display unit 2108 displays the cursor 3201. Move to the next lower category and display.
When the display unit 2108 receives “selection” from the input unit 2109, the display unit 2108 reads out the document data identification number included in the classification item indicated by the classification item name indicated by the cursor 3201, and identifies it by the read document identification number. The read document data is read from the data storage unit 2101 and the read document data is displayed.
When the display unit 2108 receives “selection” from the input unit 2109 in a state where the search menu shown in FIG. 37B is displayed, the document data identification included in the classification item indicated by the classification item name “Kinki” is displayed. The number (1) is read, and the document data 2201 identified by the read document identification number (1) is read from the data storage unit 2101 and displayed.
The display unit 2108 may display at least one set of the received classification item name and the received number of document data as a search menu.
[0098]
2.1.9 Input unit 2109
The input unit 2109 receives user input. As shown in FIG. 38, the input unit 2109 includes an up cursor key 3301, a down cursor key 3302, and a selection key 3303.
When the up cursor key 3301 is operated by the user, “upward movement” is output to the display unit 2108, and when the down cursor key 3302 is operated by the user, “downward movement” is output to the display unit 2108.
When the selection key 3303 is operated by the user, the input unit 2109 outputs “selection” to the display unit 2108.
[0099]
2.2 Search Menu Creation Operation of Search Menu Creation Device 2100
Here, the operation of creating the search menu by the search menu creating apparatus 2100 shown in FIG. 24 will be described with reference to the flowchart shown in FIG.
The keyword extraction unit 2102 extracts the keyword for each document data from the document data stored in the data storage unit 2101, and the keyword storage unit 2103 stores the extracted keyword as keyword information. (Step S3401). The associating unit 2105 extracts important word identification information added to one important word information from the important word storage unit 2103 (step S3402), and selects one of the important word information specified by the extracted important word identification information. One important word is extracted (step S3403), and the extracted important word is compared with the keyword included in the classification pattern stored in the classification pattern storage unit 2104 (step S3404). A corresponding classification item is extracted (step S3405). The associating unit 2105 repeats the comparison and extraction for all the important words extracted from the important word information specified by the extracted important word identification information (steps S3404 to S3406). The associating unit 2105 deletes the duplicated combination of the classification pattern and the classification item, and writes the important word identification information to the classification item included in the extracted classification pattern, thereby including it in the classification pattern. The classified item is associated with the document data (step S3407). Furthermore, the associating unit 2105 repeats the above comparison and association for all the document data (steps S3402 to S3408). When the association is completed for all the document data, the classification pattern selection unit 2106 calculates the evaluation value of each classification pattern using (Equation 1) (step S3409), and determines the classification pattern of the maximum evaluation value. Select (step S3410). Next, the menu creation unit 2107 creates a search menu using the classification item of the selected classification pattern as an option (step S3411), and the display unit 2108 displays the created search menu (step S3412).
[0100]
3. Other embodiments
Although the present invention has been described based on the above embodiment, it is needless to say that the present invention is not limited to the above embodiment. That is, the following cases are also included in the present invention.
[0101]
(1) In the present embodiment, after the user selects a search menu option, the input data associated with the item corresponding to the search menu option selected by the user is sequentially displayed. If there is a large amount of input data associated with the item corresponding to, a search menu may be further created based on the input data associated with the item corresponding to the selected option. For example, if the option “Kanto” is selected by the user in the search menu shown in FIG. 17, the item “Kanto” corresponding to the option “Kanto” has a plurality of small items “Tochigi” as shown in FIG. ”,“ Ibaraki ”,“ Chiba ”,“ Gunma ”,“ Saitama ”,“ Tokyo ”,“ Yamanashi ”,“ Kanagawa ”, and the menu creation unit 107 copies these multiple sub-items“ Tochigi ”. , “Ibaraki”, “Chiba”, “Gunma”, “Saitama”, “Tokyo”, “Yamanashi”, “Kanagawa”, “Chiba”, “Gunma”, “ Five sub-items “Saitama”, “Tokyo”, and “Kanagawa” are extracted, and the extracted five sub-items and the number of message board identification information included in these sub-items are output to the display unit 108 and displayed. As shown in FIG. 40, the unit 108 includes five extracted small items. The number of the message board identification information contained in these small items, displaying the cursor 1301. When the user selects one small item from the search menu displayed in this figure, the display unit 108 may sequentially display the message board information included in the selected small item.
[0102]
(2) After an option is selected from the search menu by the user, input data included in an item corresponding to the option selected by the user is associated with a classification pattern that is not selected again and newly input. A search menu may be created using the classification pattern associated with the data.
For example, in the first search menu, a hierarchical place name search menu is displayed, the user selects one place name option, and only the input data included in the item associated with the selected place name option is displayed. Again, associate with another classification pattern without a selection mark, and select one classification pattern from the classification patterns associated with these input data by the same method described in the above embodiment. Then, another search menu may be created and displayed using the selected classification pattern. In this way, a search menu with high search efficiency can be created even when a large amount of input data exists.
[0103]
(3) Depending on the number of input data associated with each classification pattern, it may be determined whether to display input data or create the next search menu again.
For example, if the number of input data associated with each classification pattern is less than or equal to a predetermined value, the input data is displayed, and if it is greater than the predetermined value, the next search menu is created again. Also good. In this way, the number of options displayed on one search menu can be set close to a predetermined value, and a search menu with higher search efficiency can be created.
[0104]
Specifically, in the search menu creation device of the above embodiment, the input unit 109 accepts selection of one option from the displayed search menu from the user, and the menu creation unit 107 further includes the input unit 109. When the selection of the user is accepted, a search menu is created using the small items included in the items corresponding to the choices accepted as selections based on predetermined criteria, or the items When it is determined whether to display the input data associated with the item, and when it is determined to create a search menu, at least one input data is associated with the item corresponding to the option for which the selection is accepted At least one small item is extracted, and a search menu using the extracted small item as an option is generated. The display unit 108 further displays a search menu generated using the small items as options when the menu creation unit 107 determines to create a search menu, and when it is determined to display input data, the display unit 108 The input data associated with the item may be read from the data storage unit 101 and the read input data may be displayed.
[0105]
Here, the predetermined criterion is that the input data is displayed if the number of input data associated with the selected classification pattern is equal to or smaller than the predetermined value, and if the number is larger than the predetermined value, the search menu is displayed. Is a criterion for generating
[0106]
Further, when the next search menu is created again, as described in (2) above, after one option is selected from the search menu by the user, it corresponds to one option selected by the user. The input data included in the item is again associated with a classification pattern without a selection mark indicating used, and a search menu is created using the classification pattern newly associated with the input data. Also good.
[0107]
The input unit 109 accepts selection of one option from the displayed search menu from the user, and the associating unit 105 further inputs data associated with an item corresponding to the option for which the selection has been accepted. Each item constituting another classification pattern excluding the classification pattern including the item corresponding to the option for which the selection is accepted, canceling all the associations between the input data and the items The first classification pattern selection unit 106 associates the selected selection data with the extracted selection data based on the predetermined criterion from the other classification patterns excluding the classification pattern including the item corresponding to the selection accepted. Then, at least one classification pattern is selected, and the menu creation unit 107 is based on the at least one classification pattern. It generates a search menu, the display unit 108 may display a search menu that is generated. For example, after the user selects the option “Kanto” displayed in the search menu, the input data associated with “Kanto” is again associated with the classification pattern other than the place name, and the above-described operation is performed from the classification pattern other than the place name. In the same manner as in the above, it is possible to select one classification pattern “price 2” and generate and display a search menu using options included in the classification pattern “price 2”.
[0108]
(4) In the first embodiment, the classification pattern may include an item including a plurality of hierarchies. For example, the detailed item may be further subordinate to the small item, and the detailed item may be subordinate to the detailed item. Correspondence unit 105 associates input data with items composed of a plurality of hierarchies, and first classification pattern selection unit 106 selects a classification pattern using items composed of a plurality of hierarchies associated with input data. Then, the menu creation unit 107 may create a search menu using the selected classification pattern.
[0109]
In addition, the menu creation unit 107 creates a search menu using items included in the selected classification pattern as options, and one option is selected by the user. Next, the menu creation unit 107 selects the selected menu. A search menu is created using the sub-items subordinate to the item corresponding to the option as an option, and one option is selected by the user. Further, the menu creating unit 107 is subordinate to the sub-item corresponding to the selected option. Create a search menu with detailed items to be selected as options. Thus, a search menu consisting of a plurality of hierarchies may be created based on items consisting of a plurality of hierarchies.
[0110]
Further, as described in (3) above, when determining whether to display input data or create the next search menu again according to the number of input data associated with each classification pattern, When it is determined that the next search menu is to be created again, a search menu consisting of a plurality of hierarchies is created as described above, and when it is determined that the input data is displayed, it corresponds to the selected option. The input data associated with the item and all items subordinate to the item may be sequentially displayed.
[0111]
In addition, the items included in the classification pattern form a hierarchical structure as described above, and the associating unit 105 stores the item corresponding to the feature point for each extracted feature point in the classification pattern storage. A search is performed from step 103, the searched item is associated with the input data including the feature point, and the input unit 109 receives selection of one option from the displayed search menu from the user. The search menu creation device further includes a determination unit, an item menu generation unit, an input data menu display, and a repetition control unit. The determination unit receives the user's selection from the input unit 109. In this case, it is determined whether to create a search menu or display input data based on a predetermined criterion, and the item menu generation unit determines that the determination unit generates a search menu. Extracting an item associated with input data from items subordinate to the item corresponding to the option accepted for selection, generating a search menu using the extracted item as an option, and the input data menu unit Is determined to display the generated search menu and display the input data when it is determined by the determination unit to create a search menu. The input corresponding to the option for which the selection is accepted and the input data associated with all the items subordinate to the item are read from the data storage unit 101, the read input data is displayed, and the repetition control is performed. Until the input data menu display unit displays the read input data, the input unit 109 accepts selection of one option, the determination by the determination unit, and the search menu by the item menu generation unit. The input unit 109, the determination unit, the item menu generation unit, and the input data menu display unit may be controlled to repeat generation and display of the search menu by the input data menu display unit.
[0112]
Further, the predetermined criterion in the determination unit is that the number of input data associated with an item corresponding to the option for which the selection is accepted and all items subordinate to the item is equal to or less than a predetermined value. If there is, input data may be displayed, and if it is larger than a predetermined value, it may be a criterion for generating a search menu.
[0113]
(5) Furthermore, a classification pattern editing unit is provided so that the user can newly register a classification pattern in the classification pattern storage unit 103 or delete a classification pattern stored in the classification pattern storage unit 103. You may do it.
In order to register the classification pattern, the input unit 109 receives an input of the classification pattern by the user, and the display unit 108 displays the classification pattern for which the input has been received. Next, the input unit 109 receives an instruction to register a classification pattern by the user, and the classification pattern editing unit registers the classification pattern for which the input has been received in the classification pattern storage unit 103.
In order to delete the classification pattern, the display unit 108 displays the classification pattern stored in the classification pattern storage unit 103, and the input unit 109 accepts designation of the classification pattern to be deleted by the user. Next, the input unit 109 receives an instruction to delete the classification pattern from the user, and the classification pattern editing unit deletes the classification pattern designated by the user from the classification pattern storage unit 103.
Here, the operations of registering and deleting classification patterns are collectively referred to as editing.
This allows you to easily register classification patterns when new classification criteria are created or new technical terms are created, and old classification patterns that are no longer used can be easily deleted. It is possible for the user to register only his / her favorite classification pattern.
[0114]
(6) Furthermore, a synonym dictionary unit comprising a phrase and its synonyms is provided, and when the associating unit 105 associates feature points of input data with items of classification patterns, feature points are extracted from the synonym dictionary unit using the feature points. And the detected feature point synonym may be associated with the classification pattern item.
In this way, by associating feature points of input data with classification pattern items via synonyms, feature points can be associated with classification pattern items similar to feature points, and input data can be associated with classification patterns. The number of cases that cannot be attached can be reduced.
[0115]
(7) Furthermore, when the switching key 1404 is pressed by the user, it is considered that the classification pattern used at that time does not match the user's preference. You may make it memorize | store in the memory | storage part 104. FIG. Specifically, when the switch key 1404 is pressed, a value obtained by subtracting 1 from the priority value of the classification pattern that has been used may be stored in the priority storage unit 104 as a new priority. .
Thus, the priority of the classification pattern can be learned according to the user's preference.
[0116]
(8) In the first embodiment, when determining the optimum classification pattern, first, the number of items is evaluated, and further, the evaluation is performed based on the priority. You may evaluate by, and may also evaluate by the number of items. Alternatively, an evaluation standard may be created by combining these two evaluations, and an optimum classification pattern may be determined using the evaluation standard created in this way.
Further, an optimal classification pattern may be determined using only the priority.
[0117]
(9) In the first embodiment, the data feature point extraction unit 102 extracts feature points from the message board information and writes them to the data storage unit 101. However, the following may be used.
The data storage unit 101 stores input data as a relational database. The relational database is a collection of input data based on a table, and the table includes at least one row and at least one column. A row is composed of a sequence of at least one value, and a column is a collection of at least one value having the same data type. One value in one column is the minimum unit of input data that can be selected from the table or updated. The data type includes a character string type, a numeric type, and the like. A column having a character string type data type is a collection of character strings, and a column having a numeric data type is a collection of numeric values.
Here, each line corresponds to the message board information 260 of the above embodiment.
The associating unit 105 extracts at least one value from the sequence of at least one value constituting one row stored in the relational database in the data storage unit 101, and characterizes the extracted at least one value. Let it be a point. In this case, the data feature point extraction unit 102 becomes unnecessary.
The purchase transfer distinction 262, the address 263, the information provider name 264, the age 265, and the occupation 266 that constitute the message board information 260 correspond to the above-mentioned relational database columns.
[0118]
(10) In the first embodiment, the message board information 260 is associated with the item 711 included in the classification pattern 701 on the basis of the message board information 260, but based on the item 711 included in the classification pattern 701. Thus, the message board information 260 and the item 711 may be associated with each other.
In this case, the associating unit 105 sequentially reads items included in each classification pattern from the classification pattern storage unit 103, sequentially reads out message board information from the data storage unit 101, and for each read message board information, the message board information Extracting at least one feature point included in the image, determining whether the read item matches the extracted feature point, and reading the item and the extracted feature point; Are determined to match, the read item is associated with the message board information including the feature point that matches the item.
[0119]
(11) In the second embodiment, the number of document data associated with the classification pattern as three evaluation elements, the number of classification item names associated with the document data, and the associated document data The evaluation value of each classification pattern is calculated using (Equation 1) including variance, and the classification pattern with the highest evaluation value is selected. However, only the number of document data associated with the classification pattern is document data. Assuming that the evaluation value of each classification pattern is calculated using an evaluation formula that includes only the number of classification item names associated with or the distribution of the associated document data, and the classification pattern with the highest evaluation value is selected. Also good. Alternatively, an evaluation value obtained by combining two of these evaluation elements may be used to calculate an evaluation value of each classification pattern and select a classification pattern.
[0120]
(12) In the first embodiment, the user selects one item from the search menu, but the user may select a plurality of items from the search menu. In this case, the input unit 109 further includes a plurality of designation keys, and when the plurality of designation keys are operated by the user while the search menu is displayed on the display unit 108, the display unit 108 is moved by the cursor 1301. The displayed item is highlighted. The highlighted item indicates that this item has been selected by the user. When the user further moves the cursor 1301 upward or downward using the up cursor key 1401 or the down cursor key 1402 to position the cursor 1301 in another item of the search menu, and the user operates the plurality of designation keys. The display unit 108 reverse-displays the item indicated by the cursor 1301. In this way, a plurality of items are selected by the user. Next, when the execution key 1403 is operated by the user, the display unit 108 reads the input data associated with the selected items from the data storage unit 101 and displays the read input data.
In addition, as described above, when a plurality of items are selected by the user, as described in (1) and (2) above, input data associated with the selected items is used. Further, a search menu may be created and displayed.
The same applies to the second embodiment.
[0121]
(13) In the first embodiment, the first classification pattern selection unit 106 and the second classification pattern selection unit 110 select one classification pattern, but the first classification pattern selection unit 106 or the second classification pattern selection unit 106 The classification pattern selection unit 110 may select a plurality of classification patterns. For example, the first classification pattern selection unit 106 may calculate the number of items associated with the input data and select a plurality of items for which a value closest to the number of appropriate options 6 is calculated. . As described above, when the first classification pattern selection unit 106 or the second classification pattern selection unit 110 selects a plurality of classification patterns, the menu creation unit 107 responds to input data from the selected plurality of classification patterns. Extract attached items and create a search menu with the extracted items as options.
The same applies to the second embodiment.
[0122]
(14) As one embodiment, a program for executing the above-described procedure by a computer is recorded on a computer-readable recording medium such as a floppy disk and transferred, or the program is communicated It may be easily implemented by another independent computer system by transferring it through a line.
[0123]
【The invention's effect】
  As described above, the present inventionA search menu creation device for creating a search menu including options used for searching for a plurality of input data, a data storage means having a storage area for storing a plurality of input data, and a plurality of search menu options Classification pattern storage means in which a plurality of classification patterns composed of the above items are stored in advance, receiving means for receiving a plurality of input data, writing means for writing the received plurality of input data to the data storage means, Association means for associating each of the plurality of input data written in the data storage means with an item included in any of the plurality of classification patterns stored in advance in the classification pattern storage means; and the classification pattern Each item constituting the classification pattern for each of a plurality of classification patterns stored in advance in the storage means Item that constitutes the selected classification pattern, and the classification pattern selection means for selecting the classification pattern based on the variance of the input data calculated from the counted value A menu extracting unit that extracts an item associated with the input data, generates a search menu using the extracted item as an option, and a display unit that displays the search menu generated by the menu extracting unit. .
  According to this configuration, a plurality of classification patterns composed of a plurality of items that can serve as search options are prepared in advance, and the number of input data associated with each item among the plurality of prepared classification patterns is as uniform as possible. It is possible to automatically select priority classification patterns and select items included in this priority classification pattern automatically as a choice to generate a search menu, so multiple written inputs This has the effect of making it possible to create a search menu that is optimal for searching data.
[0124]
  TheAnd a display unit for displaying the search menu generated by the menu extraction unit.SoThe generated search menu is displayed.AndThe user can view the generated search menu.
[0125]
  TheAnd receiving means for receiving a plurality of input data from the outside, and writing means for writing the received plurality of input data to the data storage means.SoThere is an effect that a search menu can be generated based on a plurality of input data received from the outside.
[0126]
Here, the data storage means stores a data identifier for identifying each input data, and the associating means sequentially reads the input data from the data storage means, and inputs the input data for each read input data. A feature point extraction unit for extracting feature points included in the data; an item search unit for searching for an item corresponding to the feature point from the classification pattern storage unit for each of the extracted feature points; and the searched item In addition, a data association unit for adding a data identifier for identifying the input data may be included.
According to this configuration, since the input data and the classification pattern are associated with each other using the feature points extracted from the input data, there is an effect that a menu corresponding to the feature points extracted from the input data can be created.
[0127]
Here, each input data stored in the data storage means is document data, and the feature point extraction unit includes a data reading element for sequentially reading the document data from the data storage means, and the read document data A morphological analysis element that performs morpheme analysis on each document data and decomposes it into morphemes for each document data, and extracts important words from the decomposed morphemes, and important words having a part or all of the extracted important words as feature points You may comprise so that an extraction element may be included.
According to this configuration, since the key word extracted by performing morphological analysis on the document data is used to associate the document data with the classification pattern, the feature points extracted from the non-stylized data such as the document data are used. There is an effect that a corresponding menu can be created.
[0128]
Here, the important word extraction element is an important word candidate extraction element that extracts a keyword candidate from the decomposed morpheme, and the number of times that the same keyword candidate as the extracted keyword candidate appears in the same document data. The number of appearance count elements for counting and the extracted important word candidates are extracted as important words from the extracted important word candidates, and the extracted number of important words is a predetermined number of times. You may comprise so that an extraction element may be included.
According to this configuration, since an important word candidate included a predetermined number of times or more is extracted from the document data as an important word, there is an effect that an appropriate feature point corresponding to the document data can be extracted.
[0129]
Here, the important word extraction element is a key word candidate extraction element that extracts a key word candidate from the decomposed morpheme, and the number of times that the same key word candidate as the extracted key word candidate appears in the same document data. A predetermined number of important word candidates are extracted as important words in order from the extracted important word candidates and the extracted important word candidates in order from the important word candidates with the largest number of counted times. It may be configured to include a predetermined number of extraction elements having a feature point.
According to this configuration, since a predetermined number of important word candidates are extracted as important words in order from important word candidates that are included in the document data, an appropriate feature point corresponding to the document data can be extracted. .
[0130]
Here, the data storage means stores the plurality of input data as a relational database, each input data corresponds to a row of the relational database, and is composed of an array of at least one value. A data reading element for sequentially reading input data from the data storage means and a data extracting element for extracting at least one value constituting the read input data as a feature point may be included.
According to this configuration, since each element stored in the relational database is used as a feature point as it is, there is an effect that an appropriate feature point can be quickly extracted from the input data.
[0131]
Here, the items included in the classification pattern storage unit form a hierarchical structure, and the association unit further extracts, for each extracted feature point, the item corresponding to the feature point as the classification pattern. A hierarchical item search unit that searches from storage means may be included, and the data association unit may be configured to associate the searched item with the input data including the feature point.
According to this configuration, the items constituting the classification pattern form a hierarchical structure, and there is an effect that input data can be associated with an item composed of a plurality of layers.
[0132]
Here, further, an input receiving unit that accepts selection of an option from the displayed search menu from the user, and when the input receiving unit receives the user's selection, a search menu is created based on a predetermined criterion. Determining means for determining whether or not to display input data, and, if the determination means determines that a search menu is to be created, from the items subordinate to the item corresponding to the option for which the selection is accepted An item menu generating unit that extracts an item associated with input data and generates a search menu using the extracted item as an option; and the determination unit generates the search menu when it is determined to generate the search menu. If the search menu is displayed and it is determined that the input data is displayed, the selection is accepted. And input data associated with all items subordinate to the item are read from the data storage means, and the input data menu display means for displaying the read input data, and the input read by the input data menu display means Until the data is displayed, the selection of one option by the input reception unit, the determination by the determination unit, the generation of the search menu by the item menu generation unit, and the search menu by the input data menu display unit You may comprise so that the input reception means, the determination means, the item menu generation means, and the repetition control means for controlling the input data menu display means may be provided so as to repeat the display.
According to this configuration, it is determined whether to display a search menu for items in each hierarchy or to display input data included in the items based on a predetermined criterion, so that the user can quickly search for desired input data. effective. In addition, since a multi-level hierarchical menu is realized by creating a search menu again for selected data, there is an effect that a user can easily search a large amount of data.
[0133]
Here, the predetermined criterion in the determination means is that the number of input data associated with the item corresponding to the option for which the selection has been accepted and all items subordinate to the item is equal to or less than a predetermined value. If so, the input data may be displayed, and if it is larger than a predetermined value, it may be configured to be a criterion for generating a search menu. According to this configuration, according to the number of input data associated with the items of each hierarchy and all items subordinate to the item, it is determined whether to display the search menu or to display the input data. When the selected data becomes small, it is possible to prevent an unnecessary search menu from being displayed, and the user can quickly search for desired input data.
[0134]
Here, the data storage means stores a data identifier for identifying each input data, and the associating means includes an item reading unit that sequentially reads items included in each classification pattern from the classification pattern storage means. , Sequentially reading input data from the data storage means, and extracting a feature point included in the input data for each read input data; the read item and the extracted feature point When the match determination unit determines whether or not the corresponding item is extracted and the extracted feature points correspond to each other, the read item is determined. In addition, a data association unit for adding a data identifier for identifying the input data including the feature point may be provided.
According to this configuration, since each item is associated with each input data based on the item included in the classification pattern, there is an effect that a search menu corresponding to the item included in the classification pattern can be created.
[0139]
  Here, the classification pattern selection means is configured to input the input data associated with each item for each classification pattern.A variance counter for counting the variance value of the number;The smallest variancevalueAnd a distribution selection unit that selects the calculated classification pattern.
  According to this configuration, the classification pattern selection unit distributes the number of data associated with each item for each classification pattern.valueAnd calculate the variancevalueSince the classification pattern is selected based on the menu item, it is possible to preferentially create menu choices from a classification pattern in which the number of data associated with each item is as uniform as possible, thereby creating an effective search menu.
[0140]
  Here, the classification pattern selection means includes an input data number counting unit that counts the number of types of input data associated with items for each classification pattern, and is associated with each item for each classification pattern. A classification pattern is selected based on a variance calculation unit that calculates a variance value of the input data from a count value of the number of input data, and the number of types of the input data counted and the variance value of the calculated input data And an input data number selection unit.
  According to this configuration, the classification pattern selection unit calculates the number of the associated data for each classification pattern, and selects the classification pattern based on the calculated number of data and the variance value. Menu options can be created preferentially from a classification pattern in which a lot of data is associated and the number of data associated with each item is as uniform as possible, and an effective search menu can be created.
[0141]
Here, the classification pattern selection unit includes a priority storage unit that stores a priority for each classification pattern, and reads out the priority from the priority storage unit, and uses the read priority to perform one classification. You may comprise so that the priority selection means to select a pattern may be included.
According to this configuration, the priority indicating the order of selecting the classification pattern is stored for each classification pattern, and the classification pattern is selected based on the priority. Therefore, the priority according to the user's preference can be set in advance. Thus, it is possible to create a search menu according to the user's preference.
[0142]
Here, the search menu creating device further includes a switch receiving unit that receives a switching instruction for the search menu displayed on the display unit from the user, and the switching receiving unit that receives the switching instruction for the user. A second classification pattern selection means for selecting a classification pattern based on the predetermined criteria in the classification pattern selection means from other classification patterns excluding the classification pattern selected by the classification pattern selection means; The extraction unit may be configured to generate a search menu from the classification pattern selected by the second classification pattern selection unit.
According to this configuration, when a user gives an instruction, a menu can be created by selecting another classification pattern, so that the user can switch to a favorite menu.
[0143]
Here, the search menu creation device further includes an input reception unit that receives selection of an option from the displayed search menu from the user, and the association unit further corresponds to the option for which the selection has been received. Selection input data extracting means for extracting input data associated with an item to be selected as selection data, correspondence canceling means for canceling all the correspondence between the input data and the item, and an option for accepting the selection A selection input data associating means for associating each item constituting another classification pattern excluding a classification pattern including an item corresponding to the item and the extracted selection data, wherein the classification pattern selection means performs the selection. Based on the predetermined criteria from other classification patterns excluding the classification pattern including the item corresponding to the accepted option. There are, select a classification pattern, the menu extraction means, based on the classification pattern, generates a search menu, the display unit may be configured to display a search menu that is generated.
According to this configuration, the input data associated with the option selected by the user is associated with another classification pattern again, and a search menu is created again. Thus, the user can quickly search for desired input data.
[0144]
Here, a classification pattern input accepting unit that accepts input of a classification pattern and a classification pattern editing instruction including registration of the classification pattern and deletion of the classification pattern from the user, and an input by the classification pattern input acceptance unit A classification pattern editing unit that edits a classification pattern stored in the classification pattern storage unit based on the classification pattern and the classification pattern editing instruction may be included.
According to this configuration, since the user can freely register or delete the classification pattern, there is an effect that a favorite menu can be displayed.
[0145]
  The present invention also provides a plurality of input data.Has a storage area for storingClassification pattern consisting of data storage means and multiple itemsIn advanceA plurality of classification pattern storage means for storingTo operate as a search menu creation device for creating a search menu including options for using a computer to search a plurality of input dataA search menu creation method,A receiving step for receiving a plurality of input data, a writing step for writing the received plurality of input data into the data storage means, and a plurality of input data written in the data storage means, respectively, Corresponding to an item included in any of a plurality of classification patterns stored in advance in the meansA step with correspondence;SaidFor classification pattern storage meansFor each of a plurality of pre-stored classification patterns, count the number of input data associated with each item constituting the classification pattern, based on the variance of the input data calculated from the counted value,The input data includes a classification pattern selection step for selecting a classification pattern, and items constituting the selected classification pattern.InThe method includes a menu extracting step of extracting a correlated item, generating a search menu using the extracted item as an option, and a display step of displaying the search menu generated by the menu extracting unit.
  When this method is used, a plurality of classification patterns composed of a plurality of items that can be search options are prepared in advance, and the number of input data associated with each item among the plurality of prepared classification patterns is as uniform as possible. It is possible to automatically select priority classification patterns and select items included in this priority classification pattern automatically as a choice to generate a search menu, so multiple written inputs This has the effect of making it possible to create a search menu that is optimal for searching data.
[0146]
  TheAnd a display step for displaying the search menu generated by the menu extraction step.SoThe generated search menu is displayed.AndThe user can view the generated search menu.
[0147]
  TheAnd a receiving step for receiving a plurality of input data from the outside, and a writing step for writing the received plurality of input data to the data storage means.BecauseThere is an effect that a search menu can be generated based on a plurality of input data received from the outside.
[0148]
Here, the data storage means stores a data identifier for identifying each input data, and the associating step sequentially reads the input data from the data storage means, and inputs the input data for each read input data. A feature point extracting step for extracting a feature point included in the data; an item search step for searching for an item corresponding to the feature point for each extracted feature point from the classification pattern storage unit; and the searched item And a data association step for adding a data identifier for identifying the input data.
When this method is used, the input data and the classification pattern are associated with each other using the feature points extracted from the input data, so that a menu corresponding to the feature points extracted from the input data can be created.
[0149]
Here, each input data stored in the data storage means is document data, and the feature point extraction step includes a data read step for sequentially reading the document data from the data storage means, and the read document data A morpheme analysis step that performs morpheme analysis on each document data and decomposes it into morphemes for each document data, and extracts important words from the decomposed morphemes, and important words having a part or all of the extracted important words as feature points An extraction step may be included.
When this method is used, document data and classification patterns are associated using important words extracted by performing morphological analysis on the document data, so that feature points extracted from non-stylized data such as document data are used. There is an effect that a corresponding menu can be created.
[0150]
Here, in the important word extraction step, the important word candidate extraction step of extracting the important word candidate from the decomposed morpheme, and the number of times that the same important word candidate as the extracted important word candidate appears in the same document data. From the appearance count counting step for counting and the extracted important word candidates, the important word candidates whose counted number is equal to or more than a predetermined number are extracted as important words, and the extracted important words are extracted a predetermined number of times as feature points. Steps may be included.
When this method is used, since independent words included a predetermined number of times or more are extracted from the document data as important words, there is an effect that an appropriate feature point corresponding to the document data can be extracted.
[0151]
Here, in the important word extraction step, the important word candidate extraction step of extracting the important word candidate from the decomposed morpheme, and the number of times that the same important word candidate as the extracted important word candidate appears in the same document data. Counting the number of occurrences to be counted and extracting a predetermined number of important word candidates as important words in order from the extracted important word candidates having the highest counted number, and extracting the predetermined number of important words And a predetermined number of extraction steps having a feature point as a feature point.
When this method is used, since a predetermined number of independent words are extracted as important words in order from the independent words having the highest number of contained words, it is possible to extract appropriate feature points according to the document data.
[0152]
Here, the data storage means stores the plurality of input data as a relational database, each input data corresponds to a row of the relational database, and is composed of an array of at least one value, and the feature point extraction step includes: A data reading step for sequentially reading input data from the data storage means and a data extracting step for extracting at least one value constituting the read input data as a feature point may be included.
When this method is used, since each element stored in the relational database is used as a feature point as it is, an appropriate feature point can be quickly extracted from the input data.
[0153]
Here, the items included in the classification pattern storage unit form a hierarchical structure, and the association step further includes, for each of the extracted feature points, the items corresponding to the feature points to the classification pattern. A hierarchical item search step for searching from storage means may be included, and the data association step may be configured to associate the searched item with the input data including the feature point.
When this method is used, the items constituting the classification pattern form a hierarchical structure, and there is an effect that input data can be associated with items consisting of a plurality of layers.
[0154]
Here, furthermore, an input receiving step for accepting selection of an option from the displayed search menu from the user, and when the input receiving step accepts the user's selection, a search menu is created based on a predetermined criterion. Or determining whether to display input data, and if the determination step determines that a search menu is to be created, from the items subordinate to the item corresponding to the option for which the selection is accepted An item menu generation step for generating an item associated with input data and generating a search menu using the extracted item as an option; and when the determination step determines that the search menu is to be generated, the item is generated If the search menu is displayed and the input data is determined to be displayed, the selection is accepted. An input data menu display step for reading out the input data associated with the attached option and all the items subordinate to this item from the data storage means, and displaying the read input data, and the input data menu display Until the read input data is displayed by step, selection of one option by the input reception step, determination by the determination step, generation of a search menu by the item menu generation step, and the input data menu The input reception step, the determination step, the item menu generation step, and a repetition control step for controlling the input data menu display step may be included so as to repeat the display of the search menu in the display step. .
If this method is used, it is determined whether to display a search menu for items in each hierarchy or to display input data included in the items based on a predetermined criterion, so that the user can quickly search for desired input data. effective. In addition, since a multi-level hierarchical menu is realized by creating a search menu again for selected data, there is an effect that a user can easily search a large amount of data.
[0155]
Here, the predetermined criterion in the determination step is that the number of input data associated with an item corresponding to the option for which the selection has been accepted and all items subordinate to the item is equal to or less than a predetermined value. If so, the input data may be displayed, and if it is larger than a predetermined value, it may be configured to be a criterion for generating a search menu.
When this method is used, it is determined whether to display the search menu or to display the input data according to the number of input data associated with the items of each hierarchy and all items subordinate to the item. When the selected data becomes small, it is possible to prevent an unnecessary search menu from being displayed, and the user can quickly search for desired input data.
[0156]
Here, the data storage means stores a data identifier for identifying each input data, and the associating step includes an item reading step for sequentially reading items included in each classification pattern from the classification pattern storage means; , Sequentially reading input data from the data storage means, extracting a feature point included in the input data for each read input data, and reading the item and the extracted feature point A match determination step for determining whether or not the corresponding item is matched, and when the match determination step determines that the read item corresponds to the extracted feature point, the read item And a data association step of adding a data identifier for identifying the input data including the feature point.
When this method is used, each item and each input data are associated with each other based on the items included in the classification pattern, so that it is possible to create a search menu corresponding to the items included in the classification pattern.
[0161]
  Here, the classification pattern selection step is a distribution of the number of input data associated with each item for each classification pattern.A variance calculation step for calculating a value;The smallest variancevalueAnd a distributed selection step of selecting the calculated classification pattern.
  When this method is used, the classification pattern selection step performs distribution of the number of data associated with each item for each classification pattern.valueAnd calculate the variancevalueSince the classification pattern is selected based on the menu item, it is possible to preferentially create menu choices from a classification pattern in which the number of data associated with each item is as uniform as possible, thereby creating an effective search menu.
[0162]
  Here, the classification pattern selection step includes an input data number counting step for counting the number of types of input data associated with the items for each classification pattern, and an association with each item for each classification pattern. A classification pattern is selected based on a variance calculation step of calculating a variance value of the input data from a count value of the number of input data, and the number of types of the input data counted and the variance value of the calculated input data And an input data number selection step.
  According to this configuration, the classification pattern selection step calculates the number of the associated data for each classification pattern, and selects the classification pattern based on the calculated number of data and the variance value. Menu options can be created preferentially from a classification pattern in which a lot of data is associated and the number of data associated with each item is as uniform as possible, and an effective search menu can be created.
[0163]
Here, the search menu creation device further includes priority storage means for storing a priority for each classification pattern, and when a plurality of classification patterns are selected by the classification pattern selection step. For each of the plurality of classification patterns, a priority selection step of reading a priority from the priority storage unit and selecting one classification pattern using the read priority may be included.
If this method is used, the priority indicating the order of selecting the classification pattern is stored for each classification pattern, and the classification pattern is selected based on the priority. Therefore, the priority according to the user's preference can be set in advance. Thus, it is possible to create a search menu according to the user's preference.
[0164]
Here, the search menu creation method further includes a switching reception step for receiving a switching instruction for the search menu displayed in the display step from a user, and when the switching reception step receives the switching instruction for the user. A second classification pattern selection step of selecting a classification pattern based on the predetermined criteria in the classification pattern selection step from other classification patterns excluding the classification pattern selected in the classification pattern selection step, and the menu The extraction step may be configured to generate a search menu from the classification pattern selected in the second classification pattern selection step.
When this method is used, when a user gives an instruction, a menu can be created by selecting another classification pattern, so that the user can switch to a favorite menu.
[0165]
Here, the search menu creation method further includes an input reception step of receiving selection of an option from the displayed search menu from the user, and the association step further corresponds to the option for which the selection has been received. A selection input data extraction step for extracting input data associated with an item to be selected as selection data, a correspondence cancellation step for canceling all the associations between the input data and the item, and an option for which the selection is accepted Each of the items constituting the classification pattern other than the classification pattern including the item corresponding to the item, and a selection input data association step for associating the extracted selection data with the classification pattern selection step, Other classification patterns excluding classification patterns that include items corresponding to the choices accepted for selection Then, a classification pattern is selected based on the predetermined criteria, the menu extraction step generates a search menu based on the classification pattern, and the display step displays the generated search menu. You may comprise as follows.
Using this method, the input data associated with the item selected by the user is re-associated with another classification pattern, and a search menu is created again, so a search menu with a different viewpoint can be created, The user can quickly search for desired input data.
[0166]
Here, further, a classification pattern input receiving step for receiving input of a classification pattern from the user, and a classification pattern editing instruction including registration of the classification pattern and deletion of the classification pattern, and input by the classification pattern input reception step. And a classification pattern editing step of editing the classification pattern stored in the classification pattern storage means based on the classification pattern and the classification pattern editing instruction.
If this method is used, the user can freely register or delete the classification pattern, so that it is possible to display a favorite menu.
[0167]
Further, the present invention provides a computer-readable recording medium storing a search menu creation program that causes a computer to execute the search menu creation method described above, and causes the computer to execute the search menu creation method using the search menu creation program. As a result, it is clear that the same effect as the above-described search menu creating apparatus can be obtained.
[Brief description of the drawings]
FIG. 1 is a configuration diagram of a search menu creation apparatus as a first embodiment of the present invention.
FIG. 2 is a diagram showing a data structure of message board information stored in a data storage unit of the search menu creation device shown in FIG. 1;
FIG. 3 is a diagram showing an example of message board information stored in a data storage unit of the search menu creation device shown in FIG. 1;
4 is a diagram showing a data configuration of feature point information stored in a data storage unit of the search menu creation device shown in FIG. 1. FIG.
FIG. 5 is a diagram illustrating an example of feature points stored in a data storage unit of the search menu creation device illustrated in FIG. 1;
6 is a diagram showing a data configuration of classification pattern information stored in a classification pattern storage unit of the search menu creation device shown in FIG. 1. FIG.
FIG. 7 is a diagram illustrating an example of a classification pattern stored in a classification pattern storage unit of the search menu creation device illustrated in FIG. 1;
FIG. 8 is a diagram showing a data configuration of priority information stored in a priority storage unit of the search menu creation device shown in FIG. 1;
FIG. 9 is a diagram illustrating an example of priority information stored in a priority storage unit of the search menu creation device illustrated in FIG. 1;
FIG. 10 is a diagram illustrating an example of classification pattern information stored in a classification pattern storage unit of the search menu creation device illustrated in FIG. 1;
FIG. 11 is a diagram illustrating an example of classification pattern information stored in a classification pattern storage unit of the search menu creation device illustrated in FIG. 1;
12 is a diagram illustrating an example of classification pattern information stored in a classification pattern storage unit of the search menu creation device illustrated in FIG. 1. FIG.
FIG. 13 is a diagram showing an example of classification pattern information stored in a classification pattern storage unit of the search menu creation device shown in FIG. 1;
FIG. 14 is a diagram illustrating an example of classification pattern information stored in a classification pattern storage unit of the search menu creation device illustrated in FIG. 1;
FIG. 15 is a diagram illustrating an example of classification pattern information stored in a classification pattern storage unit of the search menu creation device illustrated in FIG. 1;
FIG. 16 is a diagram showing an example of a search menu created by the search menu creation device shown in FIG. 1;
FIG. 17 is a diagram illustrating an example of a search menu displayed by the display unit of the search menu creation device illustrated in FIG. 1;
FIG. 18 is a diagram illustrating an example of an input unit of the search menu creation apparatus illustrated in FIG. 1;
FIG. 19 is a diagram illustrating an example of data display by the display unit of the search menu creation device illustrated in FIG. 1;
FIG. 20 is a diagram illustrating an example of classification pattern information stored in a classification pattern storage unit of the search menu creation device illustrated in FIG. 1;
FIG. 21 is a diagram illustrating an example of classification pattern information stored in a classification pattern storage unit of the search menu creation device illustrated in FIG. 1;
FIG. 22 is a flowchart showing the operation of the search menu creation device shown in FIG. 1;
FIG. 23 is a flowchart showing a search menu creation operation of the search menu creation device shown in FIG. 1;
FIG. 24 is a block diagram of a search menu creation device as a second exemplary embodiment of the present invention.
FIG. 25 is a diagram illustrating an example of data stored in a data storage unit of the search menu creation device illustrated in FIG. 24;
26 is a diagram showing words obtained by morphological analysis of one piece of document data shown in FIG. 25. FIG.
FIG. 27 is a diagram illustrating the number of appearances of words obtained by dividing the single document data illustrated in FIG. 25 by morphological analysis.
FIG. 28 is a diagram showing a data structure of important word information stored in the important word storage unit of the search menu creating device shown in FIG. 24;
FIG. 29 is a diagram showing an example of important words stored in the important word storage unit of the search menu creating apparatus shown in FIG. 24;
30 is a diagram showing a data configuration of classification pattern information stored in a classification pattern storage unit of the search menu creation device shown in FIG. 24. FIG.
FIG. 31 is a diagram showing an example of classification pattern information stored in a classification pattern storage unit of the search menu creation device shown in FIG. 24;
32 is a diagram showing an example of correspondence between important words, classification patterns, and classification items stored in an association unit of the search menu creation device shown in FIG. 24. FIG.
FIG. 33 is a diagram showing an example in which overlapping correspondence is deleted from the correspondence between the classification pattern and the classification item stored in the correspondence unit of the search menu creation device shown in FIG. 24;
34 is a diagram showing an example in which an identification number of document data is added to the classification pattern stored in the classification pattern storage unit of the search menu creation device shown in FIG. 24. FIG.
FIG. 35 is a diagram showing an example in which an identification number of document data is added to the classification pattern stored in the classification pattern storage unit of the search menu creation device shown in FIG. 24;
36 (a) is a diagram showing an example of an evaluation formula used when the classification pattern selection unit of the search menu creation device shown in FIG. 24 evaluates a classification pattern.
FIG. 36B is a diagram illustrating a function of the second term of the evaluation formula used when the classification pattern selection unit of the search menu creation device illustrated in FIG. 24 evaluates the classification pattern.
37A is a diagram showing an example of a search menu created by a menu creation unit of the search menu creation device shown in FIG. 24. FIG. FIG. 37A is a diagram showing an example of search menu display by the display unit of the search menu creation device shown in FIG.
FIG. 38 is a diagram illustrating an example of an input unit of the search menu creation device illustrated in FIG. 24;
FIG. 39 is a flowchart showing the search menu creation operation of the search menu creation device shown in FIG. 24;
40 is a diagram showing an example of a search menu displayed by the display unit of the search menu creation device shown in FIG. 1. FIG.
FIG. 41 is a block diagram showing a configuration of a conventional menu search system.
42A is a diagram showing an example of data stored in the data storage unit of the menu search system shown in FIG. 41. FIG.
FIG. 42B is a diagram showing an example of a thesaurus stored in the thesaurus storage unit of the menu search system shown in FIG.
43 is a diagram showing a correspondence between data stored in the data storage unit of the menu search system shown in FIG. 41 and a thesaurus stored in the thesaurus storage unit.
44A is a diagram showing a hierarchical structure of a search menu created by the menu search system shown in FIG. 41 as an example. FIG.
FIG. 44B is a diagram showing an example of search menu display by the display unit of the menu search system shown in FIG.
FIG. 44C is a diagram showing an example of search menu display by the display unit of the menu search system shown in FIG.
[Explanation of symbols]
100 Search menu creation device
101 Data storage unit
102 Data feature point extraction unit
103 Classification pattern storage unit
104 Priority storage unit
105 Corresponding part
106 Classification pattern selection unit
107 Menu creation section
108 Display
109 Input section
110 Classification pattern selection unit
2100 Search menu creation device
2101 Data storage unit
2102 Keyword extraction unit
2103 Important word storage
2104 Classification pattern storage unit
2105 Corresponding part
2106 Classification pattern selection unit
2107 Menu creation section
2108 display unit
2109 Input section

Claims (31)

A search menu creation device for creating a search menu including options to be used for searching a plurality of input data,
Data storage means having a storage area for storing a plurality of input data;
Classification pattern storage means in which a plurality of classification patterns composed of a plurality of items that can be search menu options are stored in advance ,
Receiving means for receiving a plurality of input data;
Writing means for writing the received plurality of input data into the data storage means;
Each of the plurality of input data written to said data storage means, and associating means for associating the items included in any of the plurality of sort pattern stored in advance in the classification pattern storage means,
For each of a plurality of classification patterns previously stored in the sort pattern storing means, counts the number of input data associated with each item constituting the classification pattern, the variance of input data is calculated from the counted value Classification pattern selection means for selecting a classification pattern based on the value ,
From items constituting the selected sort pattern, and a menu extraction unit operable to extract the item was Installing corresponding to the input data, generates an extraction search menu the item was choices and,
A search menu creating apparatus comprising: display means for displaying a search menu generated by the menu extracting means . The data storage means stores a data identifier for identifying each input data,
The correspondence means is
A feature point extracting unit that sequentially reads input data from the data storage unit and extracts a feature point included in the input data for each read input data;
For each extracted feature point, an item search unit that searches the classification pattern storage means for an item corresponding to the feature point;
The search menu creating apparatus according to claim 1 , further comprising: a data association unit that adds a data identifier for identifying the input data to the searched item. Each input data stored in the data storage means is document data,
The feature point extraction unit includes:
A data reading element for sequentially reading out document data from the data storage means;
A morphological analysis element that performs morphological analysis on the read document data and decomposes the document data into morphemes, and
The search menu creation device according to claim 2 , further comprising: a key word extraction element that extracts a key word from the decomposed morpheme and has a feature point of a part or all of the extracted key word. The important word extraction element is:
An important word candidate extraction element for extracting an important word candidate from the decomposed morpheme;
In the same document data, an appearance count element that counts the number of times the same important word candidate as the extracted important word candidate appears,
An important word candidate whose extracted count is equal to or more than a predetermined number of times is extracted as an important word from the extracted important word candidates, and includes a predetermined number of extraction elements having the extracted important word as a feature point. The search menu creation device according to claim 3 . The important word extraction element is:
An important word candidate extraction element for extracting an important word candidate from the decomposed morpheme;
An occurrence count element that counts the number of occurrences of the same important word candidate as the extracted important word candidate in the same document data;
Among the extracted important word candidates, a predetermined number of important word candidates are extracted as important words in order from the important word candidates with the largest number of counts, and a predetermined number is extracted with the extracted predetermined number of important words as feature points. The search menu creation device according to claim 3 , further comprising: The data storage means stores the plurality of input data as a relational database, each input data corresponds to a row of the relational database, and is composed of a sequence of at least one value;
The feature point extraction unit includes:
A data read element for sequentially reading input data from the data storage means;
The search menu creation device according to claim 2 , further comprising: a data extraction element that extracts at least one value constituting the read input data as a feature point. The items included in the classification pattern storage unit form a hierarchical structure,
The correspondence means further includes:
For each extracted feature point, including a hierarchical item search unit that searches the classification pattern storage means for the item corresponding to the feature point,
The search menu creation device according to claim 2, wherein the data association unit associates the retrieved item with the input data including the feature point. Furthermore, an input receiving means for receiving selection of an option from the displayed search menu from the user,
A determination unit that determines whether to create a search menu or display input data based on a predetermined criterion when the input reception unit receives the user's selection;
When it is determined by the determination means to create a search menu, an item associated with input data is extracted from items subordinate to the item corresponding to the option for which the selection has been accepted, and the extracted item is Item menu generation means for generating a search menu as an option;
When the determination unit determines to create a search menu, the generated search menu is displayed, and when it is determined to display input data, an item corresponding to the option for which the selection is accepted and this item Input data menu display means for reading input data associated with all items subordinate to the data storage means, and displaying the read input data;
Until the input data read by the input data menu display means, the selection of one option by the input reception means, the determination by the determination means, and the generation of the search menu by the item menu generation means, , to repeat the display of the search menu by the input data menu display means, said input receiving means, said determining means, said item menu generating means, Bei example and repeat control means for controlling said input data menu display means,
The predetermined criterion in the determination means is that the number of input data associated with an item corresponding to the option for which the selection has been accepted and all items subordinate to the item is equal to or less than a predetermined value 8. The search menu creation device according to claim 7 , wherein the input data is displayed, and if it is larger than a predetermined value, it is a criterion for generating a search menu. The data storage means stores a data identifier for identifying each input data,
The correspondence means is
An item reading unit for sequentially reading items included in each classification pattern from the classification pattern storage unit;
A feature point extracting unit that sequentially reads input data from the data storage unit and extracts a feature point included in the input data for each read input data;
A match determination unit that determines whether or not the read item corresponds to the extracted feature point;
A data identifier for identifying the input data including the feature point in the read item when the coincidence determination unit determines that the read item corresponds to the extracted feature point The search menu creation device according to claim 1, further comprising: a data association unit for adding The classification pattern selection means includes:
A variance calculation unit that calculates a variance value of the input data from a count value of the input data associated with each item for each classification pattern ;
The search menu creation device according to claim 1 , further comprising: a variance selection unit that selects a classification pattern for which the smallest variance value is calculated. The classification pattern selection means includes:
An input data number counting unit for counting the number of types of input data associated with items for each classification pattern;
A variance calculation unit that calculates a variance value of the input data from a count value of the number of input data associated with each item for each classification pattern;
Based on the number of types of the counted input data and the calculated variance value of the input data, an input data number selection unit that selects a classification pattern;
The search menu creating apparatus according to claim 1, wherein: Furthermore, priority storage means for storing the priority for each classification pattern;
When a plurality of classification patterns are selected by the distribution selection unit, the priority is read from the priority storage unit for each of the plurality of classification patterns, and one classification pattern is selected using the read priority. The search menu creation device according to any one of claims 1, 10, and 11, further comprising: a priority selection unit. The search menu creation device further includes:
Switching accepting means for accepting an instruction to switch the search menu displayed on the display means from the user;
When the switching accepting unit accepts the user's switching instruction, the classification pattern is selected based on the predetermined criterion in the classification pattern selecting unit, from other classification patterns excluding the classification pattern selected by the classification pattern selecting unit. Second classification pattern selection means for selecting a pattern,
The search menu creation device according to claim 1 , wherein the menu extraction unit generates a search menu from the classification pattern selected by the second classification pattern selection unit. The search menu creation device further includes:
From the user, including an input receiving means for receiving selection of options from the displayed search menu,
The correspondence means further includes:
Selection input data extraction means for extracting, as selection data, input data associated with an item corresponding to the option for which the selection has been accepted;
Correspondence canceling means for canceling all the correspondence between the input data and the item,
Each of the items constituting a classification pattern other than a classification pattern including an item corresponding to the option for which the selection has been accepted, and selection input data association means for associating the extracted selection data;
The classification pattern selection means selects a classification pattern based on the predetermined criterion from other classification patterns excluding a classification pattern including an item corresponding to the option for which the selection has been accepted,
The menu extraction means generates a search menu based on the classification pattern,
The search menu creating apparatus according to claim 1 , wherein the display unit displays the generated search menu. Further, a classification pattern input receiving means for receiving input of a classification pattern from the user, and a classification pattern editing instruction including registration of the classification pattern and deletion of the classification pattern;
A classification pattern editing unit that edits a classification pattern stored in the classification pattern storage unit based on a classification pattern received by the classification pattern input reception unit and a classification pattern editing instruction. The search menu creation device according to claim 1 . A computer comprising a data storage means having a storage area for storing a plurality of input data and a classification pattern storage means for storing a plurality of classification patterns composed of a plurality of items in advance for searching for a plurality of input data A search menu creation method for operating as a search menu creation device for creating a search menu including options to be used ,
A receiving step of receiving a plurality of input data by a receiving means;
A writing step of writing the received plurality of input data into the data storage unit by a writing unit;
An associating step of associating each of the plurality of input data written in the data storage unit with an item included in any of the plurality of classification patterns stored in advance in the classification pattern storage unit by the associating unit; ,
Calculated by the classification pattern selection means, for each of the plurality of sort pattern stored in advance in the classification pattern storage unit, the sort pattern by counting the number of input data associated with each item constituting, from the counted value A classification pattern selection step for selecting a classification pattern based on a variance value of input data to be input ,
The menu extraction means, from the items constituting the selected sort pattern, and a menu extraction step of the extracted items digits with corresponding to the input data, generates an extraction search menu the item was choices and,
And a display step for displaying the search menu generated by the menu extraction means by a display means . The data storage means stores a data identifier for identifying each input data,
The correspondence step includes
A feature point extracting step of sequentially reading out input data from the data storage means and extracting a feature point included in the input data for each read input data;
For each of the extracted feature points, an item search step of searching for items corresponding to the feature points from the classification pattern storage unit;
The search menu creation method according to claim 16 , further comprising: a data association step of adding a data identifier for identifying the input data to the retrieved item. Each input data stored in the data storage means is document data,
The feature point extraction step includes:
A data reading step for sequentially reading out document data from the data storage means;
A morphological analysis step for performing morphological analysis on the read document data and decomposing the document data into morphemes for each document data;
The search menu creation method according to claim 17 , further comprising: extracting a key word from the decomposed morpheme and extracting a key word using a part or all of the extracted key word as a feature point. The important word extraction step includes:
An important word candidate extracting step of extracting an important word candidate from the decomposed morpheme;
In the same document data, an appearance count counting step for counting the number of times the same important word candidate as the extracted important word candidate appears,
Extracting from the extracted important word candidates important word candidates with the counted number of times equal to or greater than a predetermined number as important words, and a predetermined number of extraction steps using the extracted important words as feature points. The search menu creation method according to claim 18 . The important word extraction step includes:
An important word candidate extracting step of extracting an important word candidate from the decomposed morpheme;
In the same document data, an appearance count counting step for counting the number of times the same important word candidate as the extracted important word candidate appears,
Among the extracted important word candidates, a predetermined number of important word candidates are extracted as important words in order from the important word candidates with the largest number of counts, and a predetermined number is extracted with the extracted predetermined number of important words as feature points. The search menu creation method according to claim 18 , further comprising: The data storage means stores the plurality of input data as a relational database, each input data corresponds to a row of the relational database, and is composed of a sequence of at least one value;
The feature point extraction step includes:
A data reading step for sequentially reading input data from the data storage means;
The search menu creation method according to claim 17 , further comprising: a data extraction step of extracting at least one value constituting the read input data as a feature point. The items included in the classification pattern storage unit form a hierarchical structure,
The correspondence step further includes:
For each of the extracted feature points, a hierarchical item search step of searching the item corresponding to the feature points from the classification pattern storage means,
The search menu creation method according to claim 17, wherein the data association step associates the retrieved item with the input data including the feature point. Furthermore, an input receiving step for receiving selection of an option from the displayed search menu from the user,
A determination step for determining whether to create a search menu or display input data based on a predetermined criterion when the input reception step receives the user's selection;
When it is determined that the search menu is created by the determination step, an item associated with input data is extracted from items subordinate to the item corresponding to the option for which the selection is accepted, and the extracted item is An item menu generation step for generating a search menu as an option;
When it is determined that the search menu is created in the determination step, the generated search menu is displayed, and when it is determined that the input data is displayed, the item corresponding to the option for which the selection is accepted and the item An input data menu display step for reading input data associated with all items subordinate to the data storage means and displaying the read input data;
Until the read input data is displayed by the input data menu display step, the selection of one option by the input reception step, the determination by the determination step, the generation of the search menu by the item menu generation step, , to repeat the display of the search menu by the input data menu display step, the input receiving step, the determination step, the item menu generating step, seen including a repeat control step of controlling the input data menu display step,
The predetermined criterion in the determination step is that the number of input data associated with an item corresponding to the option for which the selection has been accepted and all items subordinate to the item is equal to or less than a predetermined value. 23. The search menu creation method according to claim 22 , wherein the input data is displayed, and if the input data is larger than a predetermined value, it is a criterion for generating a search menu. The data storage means stores a data identifier for identifying each input data,
The correspondence step includes
An item reading step for sequentially reading items included in each classification pattern from the classification pattern storage means;
A feature point extracting step of sequentially reading out input data from the data storage means and extracting a feature point included in the input data for each read input data;
A match determination step for determining whether or not the read item corresponds to the extracted feature point;
A data identifier for identifying the input data including the feature point in the read item when it is determined by the matching determination step that the read item corresponds to the extracted feature point The search menu creation method according to claim 16 , further comprising: a data association step of adding. The classification pattern selection step includes:
A variance calculating step for calculating a variance value of the input data from a count value of the input data associated with each item for each classification pattern ;
The search menu creation method according to claim 21 , further comprising: a variance selection step of selecting a classification pattern for which the smallest variance value is calculated. The classification pattern selection step includes:
An input data number counting step for counting the number of types of input data associated with items for each classification pattern;
A variance calculating step for calculating a variance value of the input data from a count value of the input data associated with each item for each classification pattern;
Based on the counted number of types of input data and the calculated variance of the input data, an input data number selection step for selecting a classification pattern;
The search menu creation method according to claim 25 , further comprising: The search menu creation device further includes priority storage means for storing a priority for each classification pattern,
Further, when a plurality of classification patterns are selected by the distribution selection step, the priority is read from the priority storage unit for each of the plurality of classification patterns, and one classification pattern is used using the read priority. 27. The search menu creation method according to claim 16 , further comprising a priority selection step of selecting. The search menu creation method further includes:
From the user, a switching reception step for receiving a switching instruction for the search menu displayed in the display step,
When the switching acceptance step accepts the user's switching instruction, the classification pattern other than the classification pattern selected by the classification pattern selection step is classified based on the predetermined criterion in the classification pattern selection step. A second classification pattern selection step of selecting a pattern,
The search menu creation method according to claim 16 , wherein the menu extraction step generates a search menu from the classification pattern selected by the second classification pattern selection step. The search menu creation method further includes:
From the user, including an input receiving step of receiving selection of an option from the displayed search menu,
The correspondence step further includes:
A selection input data extraction step for extracting, as selection data, input data associated with an item corresponding to the option for which the selection has been accepted;
A correspondence canceling step for canceling all the correspondences between the input data and the item;
A step of associating each item constituting a classification pattern other than a classification pattern including an item corresponding to an option accepted for selection with a selection input data correspondence step for associating the extracted selection data;
The classification pattern selection step selects a classification pattern based on the predetermined criterion from other classification patterns excluding a classification pattern including an item corresponding to the option for which the selection has been accepted,
The menu extraction step generates a search menu based on the classification pattern,
The search menu creation method according to claim 16 , wherein the display step displays the generated search menu. Further, a classification pattern input receiving step for receiving an input of a classification pattern from the user, and a classification pattern editing instruction including classification pattern registration and classification pattern deletion,
A classification pattern editing step for editing the classification pattern stored in the classification pattern storage means based on the classification pattern received by the classification pattern input reception step and the instruction for editing the classification pattern;
The search menu creation method according to claim 21 , further comprising: A computer-readable medium for recording a program for causing a computer to execute a search menu creation method for creating a search menu based on a plurality of input data,
A recording medium comprising a program for causing a computer to execute the search menu creation method according to any one of claims 16 to 30 .

Images