JP2005085109A - Information retrieval device and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To enable a user to output a list of useful information. <P>SOLUTION: An information retrieval part 33 retrieves relevant information in a database based on a retrieval key and a date condition inputted by an input part 32. An information similarity calculation part 34 calculates information similarity that indicates similarity to the retrieval key for each piece of information retrieved. An importance of date calculation part 35 calculates an importance level of a date that indicates a level of importance from the view point of dates according to the date condition for each piece of information retrieved. An importance of information calculation part 36 calculates a level of importance of information that indicates a level of importance as information, using the calculated information similarity and importance of date for each piece of information retrieved. A list creating part 37 makes a list that arranges items of information in order of importance levels of information. The created list is output from an output part 39. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an information search apparatus and program for searching for information in a database stored in a storage medium.

  2. Description of the Related Art Conventionally, a technique for searching for relevant information from information such as document data stored in a database based on a search key (natural language or the like) designated through a screen or the like is well known.

  Further, for example, Patent Document 1 discloses a method of displaying evaluation information (information that is useful or useless) that serves as an index for determining the degree of usefulness of searched information.

In particular, a technique for calculating a similarity to a search key for each searched information and displaying a list in which items of each information are arranged in descending order of similarity is well known. There is also known a method of displaying a list in which items of information are arranged in date order (new order).
Japanese Patent Application No. 2000-227457

  If the technique of patent document 1 is used, the user can recognize the degree of usefulness regarding each searched information. However, the user must determine the importance of each piece of information by checking the evaluation information for each piece of information, and the information useful for the user cannot be viewed as a list.

  On the other hand, in a method of displaying a list in which items of information are arranged in descending order of similarity, information similar to a search key can be preferentially viewed. However, the displayed list does not consider the importance from the viewpoint of date, and does not necessarily display information useful for the user.

  Further, in the method of displaying a list in which items of information are arranged in order of date, the user can preferentially see new information. However, what the user wants is not necessarily new information. In addition, the displayed list does not consider the degree of similarity, and does not necessarily display useful information for the user.

  The present invention has been made in view of the above circumstances, and an object thereof is to provide an information search apparatus and program capable of outputting a list of information useful for a user.

  An information search apparatus according to the present invention is an information search apparatus for searching for information in a database stored in a storage medium, and is an input for inputting a search key used for information search and a date condition indicating a condition related to a date. Means, information search means for searching for relevant information in the database based on the search key and the date condition input by the input means, and individual information searched by the information search means An information similarity calculating means for calculating an information similarity indicating the similarity to the search key, and a date importance indicating the importance from the viewpoint of date for each piece of information searched by the information searching means. Date importance calculation means for calculating according to the date condition, and information similarity calculated by the information similarity calculation means for individual information searched by the information search means Information importance calculation means for calculating information importance indicating the importance as information using the date importance calculated by the date importance calculation means, and information importance calculated by the information importance calculation means Output means for outputting a list in which items of each information are arranged in descending order.

  A program according to the present invention is a program for searching for information in a database stored in a storage medium, and a computer is provided with a search key used for information search and a date condition indicating a condition related to a date. A function to input, a function to search corresponding information in the database based on the input search key and the date condition, and similarity to the search key for each searched information A function for calculating an information similarity indicating a degree, a function for calculating a date importance indicating an importance from the viewpoint of a date according to the date condition, and a function for calculating each of the searched individual information. A function for calculating information importance indicating the importance as information using the calculated information similarity and date importance for the information, and the calculated information importance in descending order. Characterized in that to realize the function of outputting a list obtained by arranging items of each information.

  A list of information useful to the user can be output.

  Embodiments of the present invention will be described below with reference to the drawings.

  FIG. 1 is a block diagram showing a configuration of a computer system according to an embodiment of the present invention. The system shown in FIG. 1 has a configuration in which a plurality of computers 1, 2, 3, 4,. Each computer is, for example, a personal computer, and the network 10 is, for example, a LAN (Local Area Network).

  The computer 1 uses an information search program to be described later, and information (for example, document data) in a database stored in a built-in storage medium or a database stored in each storage medium of another computer 2, 3, 4,. ) Is implemented.

  The computer 1 includes a CPU (Central Processing Unit) 11, a main memory 12, an HDD (Hard Disk Drive) 13, an input device 14, a network controller 15, a port 16, a display device 17, and the like.

  The CPU 11 is a processor that controls the overall operation of the computer 1, and executes various programs on the main memory 12, for example.

  The main memory 12 is used as a work area of the CPU 11, an OS (Operating System) 21 loaded from the HDD 13, an information search program 22 for performing information search, and a user input at the time of a search request (or read from the HDD 13) ) Search condition information (search key, date condition, etc.) 23 is stored.

  The HDD 13 stores various programs 24 such as an OS and applications, search condition history information 25 that is a history of search conditions input by a user at the time of a search request, and the like.

  The input device 14 corresponds to a keyboard and a pointing device, and is used for a user to request display of an information search screen or to input various information (search key, date condition, etc.) to make a search request. used.

  The network controller 15 is, for example, a LAN controller, and controls communication between the computer 1 and other computers on the network 10. The port 16 is a LAN port provided on the back of the computer main body, for example, and corresponds to a connection unit that connects the computer 1 and the network 10.

  The display device 17 is, for example, a built-in or external display, and displays various screens (such as an information search screen).

FIG. 2 is a block diagram showing a functional configuration of the information search program 22 on the main memory 12 shown in FIG. As shown in the figure, the information search program 22 includes a control unit 31, an input unit (input interface unit) 32, an information search unit 33, an information similarity calculation unit 34, a date importance calculation unit 35, and an information importance calculation. Various functions such as a unit 36, a list creation unit 37, a history management unit 38, and an output unit (output interface unit) 39 are provided.
When there is a request to display the information search screen through the input unit 32, the control unit 31 outputs the corresponding information search screen data through the output unit 39. In particular, when there is a search request with a search condition through the input unit 32, the control unit 31 generates a corresponding list by controlling the various processing units 33 to 38, and uses the data of the list as a search result. Or output through the output unit 39.

  The input unit 32 inputs a display request for an information search screen or a search request with a search condition, which is performed by the user using the input device 14 (see FIG. 1), and passes them to the control unit 31. The search condition includes a “search key” (natural language, a logical expression using a combination of multiple keywords, etc.) used for information search, and a “date condition” indicating a date-related condition (documents up to one week before are important. Etc.).

  The information search unit 33 searches for corresponding information (for example, document data) in the databases 41, 42, 43,... Based on the search key and the date condition input by the input unit 32, for example. . These databases 41, 42, 43... Are respectively stored in storage media of computers 2, 3, 4... (See FIG. 1). In addition to the databases 41, 42, 43..., A database stored in the HDD 13 of the computer 1 may be included in the search target.

  The information similarity calculation unit 34 calculates “information similarity” indicating the similarity with the search key for each piece of information searched by the information search unit 33. This information similarity is represented by “si” in the calculation. Note that “i” corresponds to an identification number assigned to each document data.

  The date importance calculation unit 35 calculates “date importance” indicating importance from the viewpoint of date for each piece of information searched by the information search unit 33 according to the date condition. The date importance is expressed by “ti” and is expressed as a function of a difference (ΔTi) between a specified date (for example, the current date or a specific date) and a date attached to the document data as follows: .

t i = f (ΔTi) (1)
The information importance degree calculation unit 36 calculates the information similarity degree si calculated by the information similarity degree calculation unit 34 and the date importance degree calculated by the date importance degree calculation unit 35 for each piece of information retrieved by the information retrieval unit 33. Using “ti”, “information importance” indicating importance as information is calculated. This information importance is represented by “Si” and is expressed as follows as a function of the information similarity si and the date importance ti.

Si = A si + B ti (2)
Or Si = C si × ti (3)
However, A, B, and C are coefficients.

  The list creation unit 37 is calculated in a specified order (for example, calculated by the information importance calculation unit 36 based on the calculation result of the information similarity calculation unit 34, the date importance calculation unit 35, or the information importance calculation unit 36). A list is created in which items of information are arranged in descending order of information importance.

  The history management unit 38 stores the search key, date condition, search date, and the like input through the input unit 32 as history information in the HDD 13 (see FIG. 1), and reads them when requested. is there. In addition to the search key and date condition, the search result by the information search unit 33 and the various calculation results by the various calculation units 34 to 36 may be stored together as history information or read out. .

  The output unit 39 outputs the requested information search screen data, or outputs the data of the list created by the list creation unit 37 as a search result. The data output from the output unit 39 is displayed on the screen of the display device 17 (see FIG. 1).

The control unit 31 further has the following functions.
For example, when the user specifies a calculation method (calculation formula, function, etc.) used for calculating the date importance on the information search screen, the control unit 31 applies the calculation method to the date importance calculation. The date importance calculation unit 35 can be instructed to do so.

  Further, depending on the situation (for example, when the history information is specified to be used), the control unit 31 determines the date based on the history information on the storage medium managed by the history management unit 38. It is possible to determine a calculation method (calculation formula, function, etc.) used for calculating the importance and instruct the date importance calculation unit 35 to apply the calculation method to the calculation of the date importance.

  Further, the control unit 31 determines a calculation method used for calculating the date importance based on the relationship between the date condition input in the past and the corresponding search result, and applies the calculation method to the calculation of the date importance. The date importance calculation unit 35 can be instructed to do so.

  Further, according to the situation, the control unit 31 determines that the calculation method used for calculating the date importance is also determined using the information similarity calculated by the information similarity calculation unit 34. The unit 35 can be instructed.

FIG. 3 is a diagram illustrating an example of an information search screen displayed on the display device 17.
In the example of the information search screen, an input field for inputting a “search key” and an input field for inputting a “date condition” are provided. In the input field of “search key”, for example, a natural language such as “about a message from A” is input. On the other hand, in the “date condition” input field, for example, a condition such as “importance of documents up to one week ago” is selected from a plurality of candidates prepared in advance and determined. In addition, it is configured so that a specific date can be input to specify “important documents on a specific date”, or the first candidate and the second candidate can be specified as specific dates. Also good.

The user inputs necessary items in each input field and then presses a “search document” button to execute a search request. This search request is sent to the information search program 22.

4 is a diagram showing a modification of the information search screen shown in FIG.
In this example of the information search screen, it is possible to input “search formula for calculating date importance” in the date condition input field. For example, the user can specify a specific date importance calculation method (calculation formula, function, etc.) by selecting from a plurality of prepared candidates or newly inputting it. .

Next, referring to FIG. 5 to FIG. 8, the “date condition” specified on the information search screen and the “date importance” calculated by the date importance calculation unit 35 (see FIG. 2). The relationship will be described.
FIG. 5 is a diagram showing the relationship between the date importance and the date of the document data when “Importance of documents up to one week ago” is designated as the date condition. In this case, the same date importance is given to documents dated one week before the search date. On the other hand, for documents with a date of more than one week, the older the date, the lower the importance of the date.

  FIG. 6 is a diagram illustrating the relationship between the date importance and the date of the document data when “Importance of documents up to one month ago” is designated as the date condition. In this case, the same date importance is given to documents dated one month before the search date. On the other hand, for documents with dates older than one month, the older the date, the lower the importance of the date.

  FIG. 7 is a diagram showing the relationship between the date importance and the date of the document data when “Importance of documents up to the previous search date” is designated as the date condition. In this case, the same date importance is given to documents with dates from the search date (current) to the previous search date. On the other hand, for documents with a date older than the previous search date, the older the date, the lower the date importance.

  FIG. 8 shows an example in which the calculation method used for calculating the date importance is determined using the information similarity calculated by the information similarity calculation unit 34. The information similarity calculation unit 34 calculates the information similarity using the “search formula for calculating the date importance” in FIG. 4 as a search condition. The sum is calculated and normalized within the weekly section of the search similarity, and this is used as the date importance. In calculating the date importance, it is possible to simply use the number of hits in the section instead of the sum of the search similarities.

  Next, various list lists that can be created by the list creation unit 37 will be described with reference to FIGS. Here, let us consider a case where the user designates “about a message from A” as the search key and “imports documents up to one month ago” as the date condition on the information search screen. Further, it is assumed that the current date for performing the search is July 30, 2003.

  FIG. 9 is a diagram showing an example of a list of document data obtained when a normal natural language search is performed. That is, in the list L1 shown in FIG. 9, the individual document data searched by the information search unit 33 is ordered in descending order of the information similarity si (score), which is the calculation result of the information similarity calculation unit 34. It is an arrangement. In this case, for example, the order of the document data with ID = 31 indicated by the underline is the third place. This list does not consider the date importance of individual document data.

  FIG. 10 is a diagram illustrating an example of a list of document data obtained when sorting is performed based on dates. That is, the list L2 shown in FIG. 10 is based on the individual document data searched by the information search unit 33 from the one with the latest date (the one closest to the current date “July 30, 2003”). They are arranged in order. In this case, the order of the document data with ID = 31 is fourth.

  FIG. 11 is a diagram showing an example of a list showing the result of calculating the date importance. In other words, the list list L3 shown in FIG. 11 is processed by the date importance calculation unit 35 with respect to the individual document data shown in the list list L1 of FIG. The result of calculating the importance ti (score) is shown.

  Here, the relationship shown in FIG. 6 is applied as the relationship between the date importance and the date of the document data. In this case, the date importance ti is expressed by the following equation, for example.

t i = f (ΔT i) = 1 (when ΔT i <= 30),
exp (−ΔTi / 200) (when ΔTi> 30) (4)
Here, ΔTi is the difference (number of days) between the current date and the date associated with the document data. One month is regarded as 30 days.

  Since the current date is July 30, 2003 and the date of the document data with ID = 31 is July 10, 2003, ΔTi of the document data is 20. That is, since ΔTi <= 30, the date importance ti of the document data with ID = 31 is “1” which is the highest value.

  FIG. 12 is a diagram illustrating an example of a list of document data obtained when recalculation is performed using date importance. That is, the list L4 shown in FIG. 12 is calculated by the information importance calculation unit 36 using the information similarity si and the date importance ti for the individual document data searched by the information search unit 33. The results are arranged in order from the highest information importance Si (score). Here, the above-described equation (3) is applied in the calculation of the information importance Si (provided that the coefficient C = 1). In this case, the information importance Si of the document data with ID = 31 is 8000, and the ranking is first.

  As described above, the arrangement order of the document data with ID = 31 is the third place (see FIG. 9) when the date importance is not considered, whereas the first order when the date importance is considered. (See FIG. 12).

  Next, the operation of the information search program 22 of this embodiment will be described with reference to the flowchart of FIG.

  When there is a search request from the user, search condition information (search key and date condition) indicated in the search request is input through the input unit 32 (step S1). Then, the information retrieval unit 33 retrieves corresponding document data in the databases 41, 42, 43,... Based on the input retrieval key and date condition (step S2). Next, the information similarity calculation unit 34 calculates “information similarity” indicating the similarity to the search key for each document data searched by the information search unit 33 (step S3).

  In addition, the date importance calculation unit 35 calculates “date importance” indicating the importance from the viewpoint of the date according to the date condition for each document data searched by the information search unit 33 (step S4). ). Then, in the information importance calculation unit 36, the information similarity calculated by the information similarity calculation unit 34 and the date calculated by the date importance calculation unit 35 for each document data searched by the information search unit 33. The “information importance” indicating the importance as information is calculated using the importance (step S5).

  In the list creation unit 37, based on the calculation result in the information importance degree calculation unit 36, a list list in which items of information are arranged in descending order of information importance degree calculated by the information importance degree calculation unit 36 is created. (Step S6).

  The created list data is output as a search result through the output unit 39 and displayed on the screen of the display device 17 (step S7).

  As described above, according to the present embodiment, not only the search key specified by the user but also the date condition is input, the information similarity indicating the similarity with the search key, and the date indicating the importance from the viewpoint of the date Using the importance level, an information importance level indicating the importance level as information is obtained. Since a list is created and displayed according to this information importance, it is possible to present a list of document data that matches the user's request.

  Note that the present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. In addition, various inventions can be formed by appropriately combining a plurality of components disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, constituent elements over different embodiments may be appropriately combined.

  For example, the various processing procedures according to the present invention described in the above-described embodiments are stored as a computer program in a computer-readable storage medium (for example, a magnetic disk, an optical disk, or a semiconductor memory), and as necessary. It may be read and executed by a computer (processor). Such a computer program can also be distributed by transmitting from one computer to another computer via a communication medium.

1 is a block diagram showing a configuration of a computer system according to an embodiment of the present invention. The block diagram which shows the function structure of the information search program on the main memory shown in FIG. The figure which shows an example of the screen for information searches displayed on a display apparatus. The figure which shows the modification of the screen for information searches shown in FIG. The figure which shows the relationship between the date importance degree and the date of document data in the case where “Importance of documents up to one week ago” is designated as the date condition. The figure which shows the relationship between the importance of a date, and the date of document data in the case where "Importance of documents up to one month ago" is specified as the date condition. The figure which shows the relationship between the date importance degree and the date of document data in the case where “Importance of documents up to the previous search date” is designated as the date condition. The figure which shows the relationship between the importance of a date, and the date of document data in the case where "Importance of a document on a specific date" is designated as the date condition. The figure which shows the example of the list of the document data obtained when performing a normal natural language search. The figure which shows the example of the list list of the document data obtained when sorting on the basis of a date. The figure which shows the example of the list which shows the result of calculating the date importance. The figure which shows the example of the list list | wrist of the document data obtained when recalculating using a date importance. The flowchart which shows the operation | movement by the information search program of this embodiment.

Explanation of symbols

  1, 2, 3, 4, ... Computer, 10 ... Network, 11 ... CPU, 12 ... Main memory, 13 ... HDD, 14 ... Input device, 15 ... Network controller, 16 ... Port, 17 ... Display device, 21 ... OS , 22 ... Information search program, 23 ... Search condition information, 24 ... Various programs, 25 ... Search condition history information, 31 ... Control part, 32 ... Input part (input interface part), 33 ... Information search part, 34 ... Information similarity Degree calculation unit, 35 ... date importance calculation unit, 36 ... information importance calculation unit, 37 ... list creation unit, 38 ... history management unit, 39 ... output unit (output interface unit), 41 ... database.

Claims (10)

An information retrieval device for retrieving information in a database stored in a storage medium,
An input means for inputting a search key used for information search and a date condition indicating a condition related to the date;
Information search means for searching for relevant information in the database based on the search key and the date condition input by the input means;
Information similarity calculation means for calculating information similarity indicating the similarity to the search key for each piece of information searched by the information search means;
Date importance calculation means for calculating the date importance indicating the importance from the viewpoint of the date according to the date condition for the individual information searched by the information search means;
For each piece of information retrieved by the information retrieval means, the information similarity is calculated using the information similarity calculated by the information similarity calculation means and the date importance calculated by the date importance calculation means. Information importance calculation means for calculating the information importance indicating the degree,
Output means for outputting a list in which items of information are arranged in descending order of information importance calculated by the information importance calculating means;
An information retrieval apparatus comprising:   2. The information retrieval apparatus according to claim 1, further comprising means for allowing a calculation method used for calculating the date importance to be specified. History information management means for storing at least the search key and the search date as history information in a predetermined storage medium;
Means for determining a calculation method to be used for calculating the date importance based on the history information stored in a predetermined storage medium by the history information management means;
The information search apparatus according to claim 1, further comprising:   The apparatus further comprises means for determining a calculation method used for calculating the date importance based on a relationship between a date condition input through the input means in the past and a search result searched by the information search means. The information search device according to any one of claims 1 to 3.   5. The method according to claim 1, further comprising means for determining a calculation method used for calculating the date importance using the information similarity calculated by the information similarity calculation means. The information search device described in 1. A program for retrieving information in a database stored in a storage medium, which is stored in a computer,
A function for inputting a search key used for information search and a date condition indicating a condition related to the date;
A function of searching for relevant information in the database based on the input search key and the date condition;
A function of calculating an information similarity indicating a similarity with the search key for the searched individual information;
A function for calculating the date importance indicating the importance from the viewpoint of the date according to the date condition for the retrieved individual information;
A function of calculating information importance indicating the importance as information, using the calculated information similarity and date importance for the retrieved individual information;
A function of outputting a list in which items of each information are arranged in descending order of the calculated information importance;
A program characterized by realizing.   The program according to claim 6, further causing a function of enabling a computer to designate a calculation method used for calculating the date importance. On the computer,
A function of storing at least the search key and the search date as history information in a predetermined storage medium;
A function for determining a calculation method to be used for calculating the date importance based on the stored history information;
The program according to claim 6 or 7, wherein the program is further realized.   The computer further realizes a function of determining a calculation method to be used for calculating the date importance based on a relationship between a date condition previously input through the input unit and a search result searched by the information search unit. The program according to any one of claims 6 to 8, wherein On the computer,
10. The function of further determining a calculation method used for calculating the date importance using the information similarity calculated by the information similarity calculation means is further realized. The program described in.

Images