JP2000268048A - Index preparing device and index preparing program storage medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To easily generate desired data by selecting at least one piece of data from a collected data group based on a prescribed first algorithm and preparing the index of a selected data group consisting of the selected data. SOLUTION: The data collecting part 320 of a retrieving server 300 is connected to the Internet 500 being a data source storing a source data group consisting of many pieces of data. Then, the part 320 collects data from the Internet 500 in accordance with an instruction via a terminal 100, and at least one piece of data is selected from in a collected data group consisting of the collected data by a data selecting part 321 according to operation or based on a prescribed first algorithm. An index preparing part 322 prepares the index of a selected data group consisting of the selected data. Thus, data desired by a user is easily obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an index creating apparatus for creating an index (search index) for searching data, and an index creating program storage medium for operating a computer as such an index creating apparatus.

[0002]

2. Description of the Related Art In recent years, computer networks such as the Internet and an intranet have been improved.
As the capabilities of computers have evolved, network users have been able to access a vast amount of data stored on many computers that make up the network. As a tool for obtaining data desired by a user from such a huge number of data, a search engine for searching data based on a search keyword or the like has been conventionally known.

[0003]

However, since the huge number of data described above contains data in various fields, the data is searched by a search engine and a list of search results is obtained. When a simple search keyword or the like is used, for example, a list including a lot of so-called garbage data belonging to a field other than a field that matches the interests and preferences of the user is obtained. For this reason, in order to find desired data from the search result list, it is necessary to perform an overall data check from the top to the bottom of the list and to perform a re-search by adding a search keyword, etc. There is a problem that the burden is large.
Further, when performing such a re-search or a search using a combination of a plurality of search keywords, a user who is familiar with data search can easily perform additional search keywords and combinations that can obtain desired data. , But a user unfamiliar with the search may not be able to come up with such additional search keywords.

[0004] For this reason, there is a long-awaited need for a technology that allows a user to easily obtain desired data. As one type of such technology, data in a field that matches the user's interests and preferences is listed in a search result list. There is proposed a technology aiming to be placed at a higher rank. This is because according to the list in which such data is listed at the top, the user can easily obtain desired data by checking only the top of the list.

[0005] By the way, as a technique for obtaining a list in which data in a field suited to the interests and preferences of the user are listed at the top, for example, a ranking search technique using statistical information on keywords as described below, Suitability feedback or the like for feeding back information from a user is known. The ranking search technique uses a T value that represents the number of times a certain search keyword is included in one piece of data.
Statistical information called F (Text Frequency) or IDF (Inverse Document Freq) indicating a large number of data including a certain search keyword
(Uency) is used, and data having a large TF based on a search keyword having a small IDF among search keywords input by a user is listed in a high rank in a list. The relevance feedback means that the data listed in the search result list is evaluated based on the user's interests and preferences, information representing the evaluation result is input, and the information is fed back to the search and the list is obtained again. This is a technique where the course is repeated.

However, in the above-described ranking search technique, the statistical information such as TF and IDF does not necessarily reflect the user's unique interests and preferences, and there is still a lot of garbage data at the top of the list. In addition, relevance feedback often requires the user to evaluate and feed back data many times before data in a field that matches the user's interests and preferences is listed at the top of the list. bad.

[0007] As described above, it is difficult to list data in a field that matches the interests and preferences of the user at the top of the list of search results. In the technology proposed in the related art, the user obtains desired data. In order to do so, much labor, knowledge and experience are required.

The present invention has been made in view of the above circumstances, and has as its object to provide an index creation device capable of creating an index from which desired data can be easily obtained.

[0009]

According to a first aspect of the present invention, there is provided an index creating apparatus which is connected to a data source storing a source data group including a large number of data, and which is stored in the data source. A data collection unit that collects data from the data group according to an operation; and one or more data from a collection data group including data collected by the data collection unit according to an operation or a predetermined first data. Data selection means for selecting based on an algorithm, and index creation means for creating an index of a selection data group composed of data selected by the data selection means are provided.

Here, the data collecting means may be a means for collecting the data itself, or may be a means for substantially collecting the data by collecting a title or an outline of the data.

According to the index creation apparatus of the present invention, a selection operation reflecting the interests and preferences of the user is performed as an operation on the data selection means, or an algorithm replacing the selection operation is used as the first algorithm. Is used, a selection data group having a high ratio of data in a field that matches the interests and preferences of the user and a low ratio of garbage data is obtained, and an index of such a selection data group is created. When a search is performed using such an index, garbage data is small in data found by the search, that is, high search accuracy can be realized. Then, the user can easily obtain desired data from such a search result, that is, high search efficiency is realized.

[0012] It is preferable that the index creation device of the present invention further comprises a data presentation means for further selecting and presenting data from the selected data group based on a predetermined second algorithm.

In the index creation apparatus provided with such data presentation means, as the second algorithm, for example, only data that the user currently desires is selected from among data in a field that matches the interests and preferences of the user. An algorithm based on stringent conditions is used, and the data is presented automatically. Thus, the user can receive and obtain desired data without performing a search.

In the index creation apparatus of the present invention, the data source stores data including a link indicating another data, and the data collection means collects the data while following the link. And information for specifying one of the data stored in the data source and the number of times of tracing the link are input, and the link specified as the number of times starting from the data specified by the information is entered. It is preferable to collect data while tracing.

For example, in the Internet or the like, a link indicating another data closely related to the data is often written in the data. Therefore, when finding one piece of data belonging to a field that matches the interests and preferences of the user, for example, by collecting a series of related data while following a link several times starting from that one piece of data,
It is possible to efficiently collect data in a field that meets the interests and preferences of the user.

Further, in the index creation apparatus of the present invention, the data source includes a means for searching for data stored therein, and receives a search key and searches for data using the search key. Preferably, the data collection means is configured to input a search key corresponding to an operation to a data source and collect data obtained by a search corresponding to the search key.

In many cases, the Internet or the like is provided with search means having a high search capability, and data can be efficiently collected by collecting data using such search means.

Further, in the index creation apparatus of the present invention, the data selection means selects data in accordance with an operation, and the characteristic of the data selected based on the data selection history by the data selection means. Learning means for learning
It is preferable to include a data determination unit that determines whether each data collected by the data collection unit is data having a spot color learned by the learning unit.

According to the index creating apparatus provided with such learning means, the user can easily select data by using the result of the judgment based on the spot color learned by the learning means as a reference, or After the spot color has been sufficiently learned, automatic data selection can be performed by using the determination based on the spot color as the above-described first algorithm.

According to the present invention, there is provided an index creation program storage medium for collecting data according to an operation from a source data group consisting of a large number of data, and data collected by the data collection means. Data selecting means for selecting one or more data from the collected data group according to an operation or based on a predetermined first algorithm, and an index of the selected data group consisting of the data selected by the data selecting means. And an index creation program including an index creation means for creating the index.

In the index creation apparatus of the present invention and the above-mentioned index creation program, the same names as the constituent elements constituting them, such as data selection means, are given. If so, refers to the combination of software and hardware that performs such actions,
In the case of an indexing program, it refers only to the part of the software that performs such an action.

[0022]

Embodiments of the present invention will be described below.

FIG. 1 is a diagram showing a data search system in which a first embodiment of an index creation device according to the present invention is incorporated.

FIG. 2 shows, by way of example, a search system constituted by two terminals 100 and 200 each comprising a computer system and a search server 300. These computers 100, 200 and 3 are shown.
00 are connected to each other via a network 400.

Each computer 100, 200, 300
Is a main unit 101, 201, 301 in which a CPU, a main storage device, a magnetic disk, a communication board and the like are built, and a display screen 10 in response to an instruction from the main unit 101, 201, 301
2a, 202a, 302a, display units 102, 202, 302 for displaying images and character strings,
Keyboard 10 for inputting instructions to 200 and 300
3, 203, 303, display screens 102a, 202a, 3
The mouse 104, 204, or 304 is provided for inputting an instruction corresponding to an icon or the like displayed at the position at the time of specifying an arbitrary position on the 02a. However, the display unit and the keyboard of the search server (computer) 300 are used only by the administrator of the search system, and ordinary users of the search system can use the display unit and the keyboard of the terminals (computers) 100 and 200. Interacts with the search server 300 via.

Main units 101 and 20 of terminals 100 and 200
1 also has a floppy disk, CDRO
FD loading ports 101a and 20 into which M (not shown) is loaded.
1a: CDROM loading ports 101b, 201b are provided, and inside these loading ports 101a,
101b; access by driving a floppy disk or CDROM loaded from 201a, 201b;
A floppy disk driver and a CDROM driver are also built-in. Also, in the main body 301 of the search server,
An FD loading port 301a and an MO loading port 301c into which a floppy disk and an MO (magneto-optical disk) are loaded are provided.
a, floppy disk or MO loaded from 301c
It also has a floppy disk driver and MO driver for accessing the drive.

The index creation program referred to in the present invention is an FD
Loading ports 101a, 201a; CDROM loading port 101
b, Floppy disk loaded in 201b, CDR
OM, floppy disk, MO loaded in FD loading port 301a, MO loading port 301c, network 4
From 00, it is installed in the magnetic disk of the search server 300. When the installed index creation program is executed by the CPU, the search server 300 operates as one embodiment of the index creation device of the present invention. Therefore, a floppy disk, CDROM, or MO storing an index creation program is one embodiment of the index creation program storage medium of the present invention, and a magnetic disk on which the index creation program is installed is also included in the index creation program storage medium of the present invention. It is one embodiment.

Hereinafter, the terminal 100 will be described as a representative of the two terminals 100 and 200.

FIG. 2 is a hardware configuration diagram of the search server.

The hardware configuration diagram includes a central processing unit (CPU) 311, a RAM 312, a magnetic disk controller 313, a floppy disk driver 31
4, MO driver 315, mouse controller 316,
Shown are a keyboard controller 317, a display controller 318, and a communication board 319, which are interconnected by a bus 310.

Floppy disk driver 314, MO
The driver 315 is loaded with the floppy disk 22 and the MO 20, respectively, as described with reference to FIG.
This is for accessing the loaded floppy disk 22 and MO20.

The communication board 319 is connected to the network 400 shown in FIG. This data search system can access the Internet or the like via the communication board 319 and the network 400.

FIG. 2 shows the magnetic disk 21 accessed by the magnetic disk controller 313, the mouse 304 controlled by the mouse controller 316,
A keyboard 303 controlled by a keyboard controller 317 and a CRT display (display unit) 302 controlled by a display controller 318 are also shown.

FIG. 3 is a functional block diagram of the search system shown in FIG.

In this figure, the terminal 100 and the search server 3
00 are also shown, and the Internet 500 to which they are connected via the network 400 shown in FIG. 1 is also shown. As described above, the Internet 500 stores a large amount of data, and the Internet 500 is an example of a data source according to the present invention. Further, in the search server 300, a data collection unit 320,
A data selection unit 321 and an index creation unit 322 are realized, and these constitute a first embodiment of the index creation device of the present invention.

The data collection unit 320 responds to an instruction via the terminal 100 as described below.
Collect data from 00. However, in this case, data is substantially collected by collecting data addresses, titles, and the like. Here, the method of instructing the data collection unit 320 includes a case where a directory name or a file name is specified, a case where a URL address and the number of links are specified, and a case where a search engine and a search keyword of the Internet are specified. Is arbitrarily selected by the user from the following.

FIG. 4 is a diagram showing an instruction screen on which a directory name and a file name are designated.

The instruction screen 610 is displayed on the display screen 102a shown in FIG. 1, and a box 611 is provided on the instruction screen 610, and the keyboard 1 shown in FIG.
04 or the like, a directory name or a file name is input in the box 611. As a result, a directory name or a file name is instructed to the data collection unit 320 shown in FIG. 3, and data stored in the specified directory or file is acquired by the data collection unit 320.

Further, as illustrated in FIG.
11 may include a symbol “*” representing almighty. If a directory name including the symbol “*” is input, all the corresponding directory names are entered. Data is collected from directories and files.

FIG. 5 is a diagram showing an instruction screen on which a URL address and the number of links are instructed.

The instruction screen 620 is displayed on the display screen 102a shown in FIG.
An L input box 621 and a length input box 622 are provided. A URL address is input into the URL input box 621, and the number of links is input into the length input box 622. Then, the URL collected in the URL input box 621 by the data collection unit 320 shown in FIG.
Data is collected from the address as a starting point while following the link by the number of links input to the length input box 622.

FIG. 6 is a diagram showing an instruction screen on which a search engine and a search keyword are specified.

The instruction screen 630 is displayed on the display screen 102a shown in FIG. 1. The instruction screen 630 includes a search engine name box 631 and a keyword box 63.
2 and a number box 633 are provided. When the search engine name is entered in the search engine box 631, the search keyword is entered in the keyword box 632, and the number of data is entered in the number box 633, the data collection unit 320 shown in FIG. A keyword is input, and data is acquired from the top of the list of search results by the search engine for the number of data items.

As described above, when the user's instruction is given to the data collection unit 320 through the terminal 100 shown in FIG. 3 and the data is collected from the Internet 500 by the data collection unit 320, the collection is performed as described below. A list of the selected data is displayed on the display screen of the terminal 100 by the data selection unit 321, and the data is selected by the data selection unit 321 according to a user's instruction via the terminal 100.

FIG. 7 is a view showing a selection screen on which data is displayed in a list by the data selection section 321. As shown in FIG.

The selection screen 700 displays a list including addresses 710, titles 720, and outlines 730 of each data, and a check button 740 for giving an instruction by a user is provided for each data.

Referring to the title 720 and the outline 730 displayed on the selection screen 700, the user determines whether each data belongs to a field that is in the interest or preference of the user, and determines the interest or the user's interest. The user clicks a check button 740 of data belonging to the field that matches his preference. The clicked check button 740 is highlighted. As a result, it is instructed that only the data whose check button 740 is highlighted is selected from the data displayed in the list.

Here, the judgment of the user does not need to be strict, but may be a rough judgment such as judging that all the data other than the unnecessary data is of interest or preference. Since the overwhelming part of the data stored on the Internet is unnecessary data that the user is not interested in, even with such a rough judgment, the data is sufficiently selected.

The selection screen 700 also has a function as a so-called browser, and when the user cannot judge only by the title 720 or the like, the address 710 is clicked to acquire data from the address and display the data. Is done.

As described above, when the user's instruction is given to the data selection unit 321 via the terminal 100 shown in FIG.
Data is sorted by the data sorting unit 321 according to the instruction. Based on the selected data, a search index corresponding to the index according to the present invention is created by the index creation unit 322.

The index creation unit 322 adds a file name and a U name to the data selected by the data selection unit 321.
Access is performed using the address of the RL, and a keyword or the like appearing in the data is extracted by so-called morphological analysis or the like, and a search in a so-called signature file format or inverted file format is performed using the extracted keyword or the like. Create an index.

The search index thus created is stored in the search index database 323. The above-described data collection, data selection, and index creation are repeatedly executed, and the search index stored in the search index database 323 is repeatedly updated. For example, data newly selected by the data selection unit 321 is sequentially added. Then, a search index corresponding to a data group including a large portion of data in such a field is obtained while a high percentage of data in a field that meets the interests and preferences of the user is obtained.

The search server 300 is provided with a data search unit 324 for performing data search using the search index stored in the search index database 323. A search is performed according to the instruction, and the search result is transmitted to the terminal 10.
0 is displayed on the display screen. The search results are displayed as a list using, for example, a conventional ranking search technique, and the user accesses data stored on the Internet 500 using a browser or the like based on the search results. By using the search index as described above, the data that is found as a result of the search by the data search unit 324 includes a sufficiently high ratio of data in a field that matches the interests and preferences of the user. Search accuracy and high search efficiency are realized.

The description of the first embodiment is completed above, and the second embodiment will be described below. This second embodiment is included in a search system similar to the search system in which the first embodiment is included, and redundant description will be omitted.

FIG. 8 is a functional block diagram of a search system including the second embodiment.

Among the components shown in the functional block diagram of FIG. 8, the same components as those shown in the functional block diagram of FIG. 3 are denoted by the same reference numerals, and description thereof is omitted.

FIG. 8 shows the sorting history database 32
5. Keyword learning unit 326, learning result database 3
27, a keyword extraction unit 328, a data determination unit 329,
The data presenting unit 330 is shown, and the sorting history database 325 and the like and the data collecting unit 32 described above are shown.
0, the data selection unit 321 and the index creation unit 322 constitute a second embodiment of the index creation device of the present invention.

FIG. 8 shows a sorting history database 3 in which a history of data sorting by the data sorting unit 321 is stored.
25, the sorting history database 325
As an example of a history, keywords extracted from data collected by the data collection unit 320 and extracted from data selected by the data selection unit 321 and data extracted from uncollected data among collected data are used as examples of the history. Are stored separately from the keywords.

Further, the keyword learning unit 326 determines, for example, that a keyword appearing in the selected data includes a certain keyword based on the keyword stored in the selection history database 325, Learn characteristics such as not including a combination with some other keyword. The learned spot color is stored in the learning result database 327.

The keyword extracting unit 328 extracts a keyword or the like from the data collected by the data collecting unit 320 by morphological analysis or the like. Further, the data determination unit 329 determines whether the keyword or the like extracted by the keyword extraction unit 328 satisfies the spot color stored in the learning result database 327. Here, the determination result by the data determination unit 329 is reflected on the data display by the data selection unit 321. That is, among the check buttons 740 shown in FIG.
The check button of the data determined to satisfy the spot color is highlighted. By referring to such a determination result, the user can easily specify the data to be sorted, and the burden on the user is reduced.

The data presentation unit 330 acquires the data newly selected by the data selection unit 321 by monitoring the update status of the search index stored in the search index database 323. Then, based on a condition to be satisfied by the desired data, which is input in advance by the user via the terminal 100, the newly selected data is examined, and if there is data that satisfies the condition, The data is presented to the user via the terminal 100. Such a data presentation unit 33
The user can obtain desired data that satisfies a preset condition by accessing the data presented by “0” through the terminal 100 and the Internet 500. In other words, the provision of the data presentation unit 330 allows the user to perform data search without using the data search unit 324.
Desired data can be obtained.

In the above-described embodiment, the data selection unit selects data according to the operation of the user. However, the data selection unit according to the present invention uses, for example, the data selection unit shown in FIG. Data may be automatically selected, or data may be automatically selected based on a selection criterion set in advance by a user.

[0063]

As described above, according to the index creating apparatus of the present invention, an index can be created that allows a user to easily obtain desired data.

[Brief description of the drawings]

FIG. 1 is a diagram showing a data search system in which a first embodiment of an index creation device of the present invention is incorporated.

FIG. 2 is a hardware configuration diagram of a search server.

FIG. 3 is a functional block diagram of a search system.

FIG. 4 is a diagram showing an instruction screen on which a directory name and a file name are indicated;

FIG. 5 is a diagram showing an instruction screen on which a URL address and the number of links are instructed.

FIG. 6 is a diagram showing an instruction screen on which a search engine and a search keyword are instructed.

FIG. 7 is a diagram illustrating a selection screen on which data is displayed in a list by a data selection unit.

FIG. 8 is a functional block diagram of a search system including a second embodiment.

[Explanation of symbols]

 Reference Signs List 20 MO 21 Magnetic disk 22 Floppy disk 100, 200 Terminal 300 Search server 320 Data collection unit 321 Data selection unit 322 Index creation unit (index creation means) 323 Search index database 324 Data search unit 325 Selection history database 326 Keyword learning unit 327 Learning Result database 328 Keyword extraction unit 329 Data discrimination unit 330 Data presentation unit 400 Network 500 Internet (data source)

 ────────────────────────────────────────────────── ─── Continued on the front page (72) Nobuhiro Yugami 4-1-1, Kamidadanaka, Nakahara-ku, Kawasaki City, Kanagawa Prefecture F-term within Fujitsu Limited (Reference) 5B075 KK07 ND36 NK32 NK44 NK49 NR12 PP03 PP13 PP22 PP25 PP30 PQ02 PR03

Claims (6)

[Claims] 1. A data collection means connected to a data source storing a source data group consisting of a large number of data, and collecting data according to an operation from the source data group stored in the data source. A data selection unit that selects one or more data from a collection data group including data collected by the data collection unit in accordance with an operation or based on a predetermined first algorithm; An index creation device for creating an index of a sorted data group composed of the sorted data. 2. The index creation device according to claim 1, further comprising a data presenting unit for further selecting and presenting data from the selected data group based on a predetermined second algorithm. 3. The data source stores data including a link indicating other data, wherein the data collection unit collects data while following the link. Information specifying one of the data stored in the data source and the number of times to follow the link are input, and data is collected while following the link by the number of times starting from the data specified by the information. 2. The index creation device according to claim 1, wherein the index creation device is an index creation device. 4. The data source includes means for searching for data stored therein, receiving a search key input, and searching for data using the search key. 2. The index creation device according to claim 1, wherein a search key corresponding to an operation is input to the data source, and data obtained by a search according to the search key is collected. 5. A learning means for sorting data in accordance with an operation, wherein the data sorting means learns a characteristic of the selected data based on a history of data sorting by the data sorting means. 2. The index according to claim 1, further comprising: a data determination unit that determines whether each data collected by the data collection unit is data having a spot color learned by the learning unit. Creating device. 6. A data collection means for collecting data according to an operation from a source data group consisting of a large number of data, and at least one of a collection data group consisting of data collected by said data collection means. An index comprising: data selection means for selecting data according to an operation or based on a predetermined first algorithm; and index creation means for creating an index of a selection data group composed of data selected by the data selection means. An index creation program storage medium, wherein the creation program is stored.