JP2007241889A - Information retrieval system, information retrieval method, and information retrieval program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an information retrieval system, an information retrieval method, and an information retrieval program for retrieving information matching an input condition. <P>SOLUTION: This information retrieval system is provided with an information retrieval part 13 retrieving an article including a keyword as a target article from article data 21 when a retrieval condition including the keyword is input, a link extraction part 14a extracting a record including an identifier of the target article from link data 22 and determining an identifier of the article associated with the identifier of the target article as an identifier of the link article, a point allotment extraction part 14c extracting point allotments associated with link types included in the record extracted from the link data 22 from point allotment data 18, a creator extraction part 14b extracting an identifier of a user associated with the identifier of the target article and extracting an identifier of a user associated with the identifier of the link article from the article data 21, and a score calculation part 14c summarizing the extracted point allotments for each of the identifiers of the users. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to an information search system, an information search method, and an information search program for searching for information that matches an input condition.

  In recent years, an information search technique for searching data that matches a search condition input by a user from data stored in a database or the like has been developed. In addition to searching for document information related to keywords and the like input as search conditions, search techniques such as KnowWho search for searching for users who have knowledge about search conditions have been developed.

  In this way, as a method for searching for other users having knowledge that the user wants to know, a profile including a keyword input as a search condition is searched from a plurality of profiles stored in advance. There is a technique for introducing a specified user as a user having knowledge specified by a search condition (see, for example, Patent Document 1).

Further, there is a technique for searching a keyword included in an input search condition from a profile or the like, and introducing a user who has created a large number of searched profiles as a user having knowledge specified by the search condition (for example, (See Patent Document 2)
JP 2003-271638 A JP 2003-108592 A

  However, in the conventional methods as described in Patent Documents 1 and 2 described above, there is a problem that a user who is not originally applicable is searched depending on how to write (use) the keyword in the profile.

  For example, when a message such as a bulletin board or e-mail is used as search target data, if the message is a question, even a user who does not have knowledge of using a keyword to describe a question has knowledge about the keyword. It may be searched as a user. Similarly, when a new sentence is written by quoting a sentence of a person who wrote before, even a user who does not have knowledge of quoting a keyword and describes a message is regarded as a user who has knowledge about the keyword. May be searched.

  Therefore, when a user who has created a sentence or the like including a keyword as described above is used as a search result, there is a problem that the reliability of the search result is not guaranteed and a search that meets the user's request cannot be performed.

  The present invention has been made in view of the above circumstances, and an object thereof is to provide an information search system, an information search method, and an information search program for searching for a user who meets an input condition.

  In order to solve the above problems, an information search system according to the feature of the present invention is an information search system including a user terminal that transmits a search condition and an information search apparatus that searches for information that matches the search condition. Is a storage device for storing article data having a record including an article identifier, an identifier of the user who created the article, and the body of the article, a plurality of related article identifiers, and a link for identifying the type of association A storage device that stores link data having a record associated with a type, a storage device that stores score data having a record that associates a link type and a score specified by the link type, and a search including a keyword When conditions are entered, an information search unit that searches for articles containing keywords as target articles from article data, and link data A record that includes the identifier of the target article is extracted, a link extraction unit that determines the identifier of the article associated with the identifier of the target article as the identifier of the link article, and a link that is included in the record extracted from the link data from the scoring data A score extracting unit for extracting a score associated with the type, a creator extracting unit for extracting a user identifier associated with the identifier of the target article from the article data, and extracting a user identifier associated with the identifier of the linked article; And a score calculation unit that counts the score extracted by the score extraction unit for each identifier of the user.

  According to the present invention according to the above configuration, the search accuracy can be improved because the user's score is totaled by taking into account the points set for each link type and the user who matches the condition is searched. .

  ADVANTAGE OF THE INVENTION According to this invention, the information search system, the information search method, and information search program which search the user who adapts the input conditions can be provided.

The best mode for carrying out the present invention will be described below with reference to the drawings.
FIG. 1 is a diagram for explaining the configuration of an information search system 1 according to the preferred embodiment of the present invention. The information search system 1 includes an information search device 10, an information storage device 20, and a user terminal 30.

  In this information search system 1, when a search condition input to the user terminal 30 by a user is transmitted via the network 40, the information search apparatus 10 that has received the search condition stores information that matches the search condition as an information storage device. Search from the data stored in 20. Further, the information search device 10 transmits the search result to the user terminal 30 via the network 40.

  The user terminal 30 includes an input device 31 and a display device 32. The user terminal 30 is a general computer having a central processing unit, a storage device, a memory, a communication interface, and the like (not shown). When the search condition is input via the input device 31, the user terminal 30 transmits the search condition to the information search device 10 via the communication interface. In the information search system 1 shown in FIG. 1, only one user terminal 30 is described, but the same applies even when a plurality of user terminals are connected to the network 40.

  As shown in FIG. 1, the information search device 10 includes an input / output interface 11, a keyword extraction unit 12, an information search unit 13, a score calculation unit 14, a rank determination unit 15, a transmission processing unit 16, a score setting unit 17, and score data 18. have.

  The information retrieval apparatus 10 is configured by a general computer having a central processing unit, a storage device, a memory, an input device, a communication interface, and the like. When the information search program stored in the storage device is read and executed by the central processing control device, the input / output interface 11, the keyword extraction unit 12, the information search unit 13, the score calculation unit 14, the rank determination unit 15, and the transmission The information search apparatus 10 having functions of the processing unit 16 and the score setting unit 17 is configured. In addition, the storage device of the information search device 10 stores stipulation data 18 together with the information search program. Note that the information search apparatus 10 may be configured by a single computer or may be configured by a plurality of computers.

  In the present embodiment, the data handled by the information search apparatus 10 in the search will be described as an “article” in the format shown in FIG. In the following description, an object to be searched as data that matches the search condition in the information search system 1 is the identifier of the user who created the article.

  The article shown in FIG. 2 has an “article ID” that identifies the article, “attribute information”, “title”, “text”, “classification”, and “link information”. “Attribute information” is a user ID for identifying the user who created the article, the date and time when the article was created, and the like. The “link information” includes the article ID of another article to which the article is linked and the type of link.

Here, the link of the article will be described. Figure 3 is a diagram for explaining an example of the article A ₂ is articles A ₁ and the leading article is the original article link. “Original article” refers to an article used to create a “previous article”. The link between the original article and the previous article has a “link type”, and this link type identifies how the original article was used to create the previous article.

Here, when the article A ₁ is the “target article”, the article A ₂ linked to the article A ₁ is referred to as a “link article” with respect to the target article. On the other hand, when the article A ₂ is the target article, the article A ₁ linked to the article A ₂ is referred to as a “link article” with respect to the target article.

Next, link types will be described with reference to FIG.
In FIG. 4, when the link type is “describe”, the previous article is an article created as an article explaining the contents of the original article, and the previous article can be viewed together with the original article. desirable.
If there is a relationship between the previous article with the link type "follow" and the original article, the previous article is an article created as a comment, supplement, question, etc. on the contents of the original article, and the previous article should be viewed together with the original article Is desirable.

If there is a relationship between the previous article with the link type "revise" and the original article, the previous article is an article created for the purpose of revising the original article, and it can be understood without looking at the original article when viewing the previous article can do.
If there is a relationship between the previous article with the link type "sum" and the original article, the previous article is an article created for the purpose of summarizing or summarizing the contents of the original article, and it is preferable to view the previous article rather than the original article .
If there is a relationship between the previous article and the original article with the link type “consult”, the previous article is an article referring to the contents of the original article, and it is desirable to view the original article together with the previous article.

Here, the information processing apparatus 20 will be described. The information storage device 20 is connected to the information search device 10 and stores article data 21 and link data 22.
As shown in FIG. 5, the article data 21 includes “article ID”, article attributes “user ID” and “creation date”, “title”, “text”, and “classification”. Have a record. The article ID is an article identifier. The user ID is the identifier of the user who created the article identified by the article ID, and the creation date and time is the date and time when this article was created. The title is the title of the article, the body is the body of the article, and the classification is the classification of the article. The article data 21 in FIG. 5 includes “record No.” for convenience of explanation, but this is not an essential element.

  In the record a1 of the article data 21 shown in FIG. 5, the article A is created by the user a, the title is “network”, and the classification is “II”. Further, in the record a2 of the article data 21, the article B is created by the user b, the title is “Internet”, and the classification is “II”.

  The link data 22 includes a record in which a plurality of related article IDs are associated with a “link type” that identifies a related type. The link data 22 shown in FIG. 6 identifies the “original article ID” that is the identifier of the original article, the “previous article ID” that is the identifier of the previous article with respect to the original article, and the type of link between the original article and the previous article. It has a record including “link type”. The link data 22 in FIG. 6 includes “record No.” for convenience of explanation, but this is not an essential element for the link data 22.

  The record b1 of the link data 22 shown in FIG. 6 indicates that the original article is the article D, the previous article is the article A, and the link type is “revise”. That is, according to FIG. 4, the article A is an article obtained by revising the article D. Further, the record b2 indicates that the original article is the article A, the previous article is the article E, and the link type is “consult”. That is, according to FIG. 4, the article E is an article with reference to the content of the article A.

Next, each configuration of the information search apparatus 10 will be described.
The input / output interface 11 of the information search apparatus 10 inputs data such as search condition data transmitted by the user terminal 30. Further, the input / output interface 11 transmits the result data of the information search process in the information search apparatus 10 to the user terminal 30. Data transmission / reception between the information search apparatus 10 and the user terminal 30 is performed via the network 40.

  The keyword extraction unit 12 extracts a keyword from search condition data that is a sentence including the input keyword. Specifically, the keyword extraction unit 12 extracts a keyword from a sentence based on a widely used form element analysis.

  The information search unit 13 reads the article data 21 from the information storage device 20 and searches the article data 21 for an article including a keyword extracted from the search condition data as a target article.

  The link extraction unit 14a reads the link data 22, extracts a record including the article ID of the target article from the link data 22, and determines the article ID associated with the article ID of the target article as the article ID of the link article.

  The creator extraction unit 14b reads the article data 21, extracts a user ID associated with the article ID of the target article from the article data 13, and extracts a user ID associated with the article ID of the link article.

  The scoring point extraction unit 14 c reads the scoring point data 18 and extracts the scoring points associated with the link type included in the record extracted from the link data 22 from the scoring point data 18. Specifically, when the link data 22 shown in FIG. 6 and the scoring data 18 shown in FIG. 7 are used, the scoring extraction unit 14c performs the scoring associated with the link type and the original article type, the link type, and the previous article. The points associated with the type are extracted.

  The score calculation unit 14d aggregates the points extracted by the point extraction unit 14c for each user ID. Specifically, when the user ID associated with the article ID of the target article or the article ID of the link article is extracted, the score calculation unit 14d generates score data 19 having a record that associates the user ID with the initial score. To do. Thereafter, the score calculation unit 14d adds the score extracted by the score extraction unit 14c to each score associated with the user ID of the score data 19. At this time, if the user is the creator of the original article, the score calculation unit 14d adds the score associated with the type of the original article in the score data 18 to the score associated with the user ID of the user. When the user is the creator of the previous article, the score calculation unit 14d adds the score associated with the type of the previous article in the score data 18 to the score associated with the user ID of the user.

  The rank determination unit 15 reads the score data 19 and determines a user with a high score associated with each user ID as a corresponding user. For example, the order determination device 15 rearranges the records of the score data 19 in descending order of the score, and extracts the users identified by the upper predetermined number of user IDs of the rearranged records as corresponding users.

  When a score change request including a rate of change of the score of the score data 18 is input to the input / output interface 11, the score setting unit 17 reads the score data 18 (described later with reference to FIG. 7) stored in the storage device. Then, new scoring data 18a (described later with reference to FIG. 13) in which the scoring of the scoring data 18 is changed according to the change ratio included in the change request is generated and stored in the storage device. The score extraction unit 14c reads the new score data 18a and extracts the corresponding score.

  The scoring data 18 has a record that associates the link type with the scoring specified by the link type. The scoring data 18 shown in FIG. 7 has a record that associates the “link type” with the “score” specified by either the original article type or the previous article type. The scoring data 18 in FIG. 7 includes “record No.” for convenience of explanation, but this is not an essential element. For example, when the link type is “describe”, the user who created the original article is given two points that are points associated with “describe” and “original” in the record c1. In addition, when the link type is “describe”, the user who created the previous article is given 4 points that are points associated with “describe” and “destination” in the record c2.

Next, an information search method in the information search system 1 shown in FIG. 1 will be described using the flowchart shown in FIG.
First, when the input / output interface 11 inputs the search condition data transmitted from the user terminal 30 (YES in S11), the keyword extraction unit 12 extracts a keyword used for the search from the input search condition data (S12). Specifically, when the search condition data “I want to know people who are familiar with the network” is input, the keyword extraction unit 12 extracts “network” as a keyword by general-purpose form element analysis. Thus, by extracting a keyword from a sentence, information can be searched even if the search condition input at the user terminal 30 is a sentence.

  When the keyword extraction unit 12 extracts keywords from the search condition data in step S12, the information search unit 13 reads the article data 21 from the information storage device 20 and searches for articles that match the search conditions (S13). Articles that meet the conditions are searched by searching for articles that contain the keywords extracted by the keyword extraction unit 12. When searching for an article, the information search unit 13 may search for an article that includes a keyword in the body of the article, or may search for an article that includes the keyword in any of the title, body, etc. of the article.

  Specifically, when searching for an article including the keyword “network” for the “body” of the article, the body of each record of the read article data 21 is referred to, and a record including “network” in the body is extracted. Then, the article specified by the extracted record is set as the search result of the information search unit 13. Here, as a method for obtaining the search result, the information search unit 13 extracts all articles including the keyword as a search result, extracts all articles including a specified number or more of the keywords and sets the search result, the keyword There are various methods, such as a method of extracting a prescribed number of articles from articles including a search result.

When the information search unit 13 searches for articles that match the search conditions, the link extraction unit 14a substitutes 0 for the variable n, and substitutes the number of articles searched by the information search unit 13 for the variable x (S14).
Subsequently, the link extraction unit 14a increments the value of n (S15), reads the link data 22 from the information storage device 20, and extracts the link article and link type linked to the target article from the link data 22 (S16). .

  Specifically, the link extraction unit 14a uses the article searched by the information search unit 13 as a target article, and extracts from the link data 22 a record whose original article ID or previous article ID is the article ID of the target article. Further, the link extraction unit 14a sets the article ID associated with the article ID of the target article in the extracted record as the article ID of the linked article, and sets the link type as the association between the target article and the linked article.

  Subsequently, the creator extraction unit 14b reads the article data 21 from the information storage device 20, and extracts the user ID of the user who created the article from the article data 21 (S17). Specifically, the creator extraction unit 14b extracts, from the article data 21, the user ID associated with the article ID of the article searched by the information search unit 13 as the user ID of the user who created the target article. Further, the creator extracting unit 14b extracts the user ID associated with the article ID of the article extracted by the link extracting unit 14a from the article data 21 as the user ID of the user who created the linked article.

  Thereafter, the score extraction unit 14c reads the score data 18 stored in the storage device, and extracts the score to be given to the user who created the article from the score data 18 (S18). Specifically, the score extraction unit 14 c extracts a score associated with the link type extracted by the link extraction unit 14 a from the score data 18. Here, for the user who created the original article, the score extracting unit 14c extracts the score associated with the type of the original article together with the link type, and the score extracting unit 14c is for the user who created the previous article. Then, the link type and the score associated with the type of the previous article are extracted.

Subsequently, the score calculation unit 14d adds the score extracted by the score extraction unit 14c to the score associated with the user ID in the score data 19 (S19), and updates the score data 19 (S20).
Thereafter, the score calculation unit 14d determines whether or not “n <x” is satisfied (S21). If the score calculation unit 14 determines that “n <x” (YES in S21), since the score calculation is not completed, the process returns to step S15, the link extraction unit 14a increments n (S15), and step S16. -S19 is repeated. Specifically, the case where the calculation of the score is completed is a case where a score is added to the user regarding all articles of the number x of articles searched by the information search unit 13.

  On the other hand, when the score calculation unit 14d determines that “n <x” (NO in S21), since the score calculation is completed, the ranking determination unit 15 reads the score data 19 stored in the storage device, and the corresponding user Is determined (S22). For example, the rank of the user can be determined by sorting the records of the score data 19 in descending order of the scores and extracting the users identified by the upper predetermined number of user IDs of the sorted records as the corresponding users.

  When the rank determination unit 15 determines the corresponding user rank, the transmission processing unit 16 transmits the search result data including the result to the user terminal 30 via the input / output interface 11 (S23).

Specifically, the case where the information search unit 13 searches for the article A, the article B, and the article C from the article data 21 shown in FIG. 5 will be described with reference to FIG.
First, regarding the link from the link data 22 to the article A, the link extraction unit 14a has “record b1” in which the previous article is the article A, “record b2” in which the original article is the article A, and the original article is the article A. "Record b3" and "record b4" whose original article is article A are extracted.

  Thereafter, the creator extraction unit 14b extracts the user ID “c” of the user who created the article D from the article data 21 and the user ID “a” of the user who created the article A as the creator extraction regarding the record b1. . At this time, the score calculation unit 14d generates score data 19 in which the initial value (in this case, “0 points”) of the score is associated with the extracted user ID, and stores the score data 19 in the storage device.

  Subsequently, since the link type of the record b1 is “revise”, the scoring point extraction unit 14c is associated with “revise” and “original” in the record c5 as the scoring points of the user who created the original article D. “3 points” is extracted, and “7 points” included in the record c6 in which the link type of the score data 18 is “revise” and “destination” is extracted as the score of the user who created the article A which is the previous article.

  Thereafter, the score calculation unit 14d adds “3 points” to the score associated with the user ID “c” of the creator of the “article D” that is the original article in the record b1, and the “article A” that is the previous article is added. The score data 19 including the score of each user is updated by adding “7 points” to the score associated with the user ID “a” of the creator.

  When the addition of scoring is completed for the record b1, the creator extracting unit 14b similarly extracts the user ID of the user who created the article A and the article E for the record b2. After that, when the score extraction unit 14 extracts a score for “consult”, the score calculation unit 14d adds “3 points” included in the record c9 to the score of “user a” who created “article A” as the original article. Then, “0 point” included in the record c10 is added to the score of “user b” who created “article E” which is the previous article.

  Similarly, for the record b3, the creator extracting unit 14b extracts the user ID of the user who created the article A and the article F. After that, when the scoring extraction unit 14c extracts the scoring for “consult”, the scoring calculation unit 14d adds “3 points” included in the record c9 to the scoring of the user a who created “article A” that is the original article, “0 points” included in the record c10 are added to the score of “user c” who created “article F” as the previous article.

  Further, similarly for the record b4, the creator extraction unit 14b extracts the user ID of the user who created the article A and the article G. After that, when the score extraction unit 14c extracts a score for “consult”, the score calculation unit 14d adds “3 points” included in the record c9 to the score of “user a” who created “article A” as the original article. Then, “0 point” included in the record c10 is added to the score of “user b” who created “article G” as the previous article.

  As described above, as a result of calculating the score for the article A, the score data 19 stored in the storage device has a score of “16 points” for the user “a” and a score of “0” for the user “b” as shown in FIG. The score of “point” and user c is “3 points”. Also, for article B, 21 points are added to the score of user b and 6 points are added to the score of user c by the same process, and 3 points are added to user a's score by the same process for article C. 3 points and 1 point are added to the score of user c. As shown in FIG. 10B, the score of user a is 19 points, the score of user b is 24 points, and the score of user c is 10 points. Score data 19 which is a point is stored. As a result, in the case of the scoring of the scoring data 19 shown in FIG. 10B, the ranking determining unit 15 places “user b” in the first place, “user a” in the second place, “user a”, and “user” in the third place. c) and the order.

  The user terminal 30 that has received the search result data transmitted by the transmission processing unit 16 in step S23 displays a search result screen on the display device 32. FIG. 11 is an example of the search result screen 100.

  The score data 19 shown in FIG. 10B has three records and stores the scores of three users. In the search result screen 100 shown in FIG. 11, all the users included in the score data 19 are searched. As a result. However, when the score data 19 stores the scores of a large number of users, the rank determining unit 15 determines a user of a predetermined rank (for example, the fifth rank) as a corresponding user and searches the search result screen. Then, only the top users (for example, only the top five users) may be displayed.

  In this way, in order to search for users who have knowledge about the items input as search conditions by the score calculated based on the type of each link, to search for the user who created the article containing the input items and In comparison, the search accuracy can be improved.

<Scoring setting>
Next, an example of setting the score data 18 in the score setting unit 17 will be described. In the processing described above, the score calculation unit 14 uses the standard score data 18, but it is also possible to use variable score data.

  For example, when inputting the search condition, the user can adjust the score of the score data 18 according to the user's desire. FIG. 12 is an example of a search condition input screen provided with an operation means for operating adjustment of scoring data. The search condition input screen 110 illustrated in FIG. 12 includes a search condition input unit 111, indicators 112a to 112e, and a search button 113. This search condition input screen 110 is displayed on the display device 32.

  In the search condition input unit 111, a user inputs a search condition via the input device 31. The indicators 112 a to 112 e are operated by the user to adjust the scoring data 18 via the input device 31. The search button 113 is operated by the user when pressed via the input device 31.

  When the user presses the search button 113 via the input device 31, the search condition data including the input condition input to the condition input unit 111 and the adjustment operation of the scoring data 18 by the indicators 112 a to 112 e is obtained as the information search device. 10 to send.

  For example, when the “describe” indicator 112a is operated in the direction of “not important”, the user terminal 30 performs a search condition including operation instruction data for adjusting the score given when the link type is “describe”. Data is transmitted to the information search apparatus 10. On the other hand, when the “describe” indicator 112a is operated in the direction of “important”, the user terminal 30 performs a search condition including operation instruction data for adjusting the score given when the link type is “describe”. Data is transmitted to the information search apparatus 10.

  FIG. 13 is an example of the scoring data 18a adjusted on the search condition input screen 110 shown in FIG. In the scoring data 18 described above with reference to FIG. 7, when the link type is “describe”, the user who created the original article has “2 points”, and the user who created the previous article has “4 points”. there were. Here, in the search condition input screen 110 of FIG. 12, since the indicator 112a is set so as to place importance on “describe”, the scoring data 18a adjusted on the search condition input screen 110 has a link type “describe”. ”Is set to“ 4 points ”for the user who created the original article, and“ 8 points ”for the user who created the previous article. That is, in the scoring data 18a shown in FIG. 13, the scoring given to the user when the link type is “describe” is set higher than the scoring data 18 shown in FIG.

  Further, in the scoring data 18 described above with reference to FIG. 7, when the link type is “sum”, the scoring of the user who created the original article is “3 points”, and the scoring of the user who created the previous article is “6 points”. "Met. Here, in the search condition input screen 110 of FIG. 12, since the indicator 112d is set so as not to place importance on “sum”, the scoring data 18b adjusted on the search condition input screen 110 has a link type of “sum. ", The score of the user who created the original article is set as" 1 point ", and the score of the user who created the previous article is set as" 3 points ". That is, in the scoring data 18a shown in FIG. 13, the scoring given to the user when the link type is “sum” is lower than the scoring data 18 shown in FIG.

  In the search condition input screen 110 shown in FIG. 12, the operation means for operating the adjustment of the scoring data uses the indicators 12a to 12e, but is not limited thereto. As another example, for each link type, a number between “0 (when not important)” to “100 (when important)” is input, and a predetermined value using a value according to the input number is used. A method can be considered in which the value obtained by the calculation is used as the adjusted score.

  As described above, the search result obtained using the score data obtained by using the score data adjusted by the operation from the user terminal 30 is compared with the score data obtained from the standard score data, and the user who desires the search. Search results that meet the conditions of

<Variation of point setting>
Next, another example of adjusting the score data 18 will be described.
FIG. 14 is another example of a search condition input screen provided with an operation means for operating adjustment of scoring data. A search condition input screen 120 illustrated in FIG. 14 is displayed on the display device 32 and includes a search condition input unit 121, radio buttons 122 a to 122 c, and a search button 123.

  In the search condition input unit 121, a user inputs a search condition via the input device 31. The radio buttons 122a to 122c are buttons for operating the selection of the scoring data 18b to 18d used for calculation of the score by the user via the input device 31. The search button 123 is operated by the user when pressed via the input device 31.

  Each of the radio buttons 122a to 122c is associated with an identifier of the scoring data 18b to 18d stored in the storage device. When the search button 123 is pressed after any radio button is selected by the user via the input device 31, the user terminal 30 displays search condition data including an identifier of the scoring data associated with the selected radio button. Send.

  The score setting unit 17 of the information search apparatus 10 that has received the search condition data extracts the score data identified by the identifier of the score data included in the search condition data, and the score calculation unit 14 d is extracted by the score setting unit 17. Score is calculated from the scoring data.

FIG. 15 is an example of the scoring data 18b to 18d selected and used on the search condition input screen 120 shown in FIG.
Scoring data 18 b in FIG. 15A is used when “revise” is emphasized, and is selected by the radio button 122 a on the search condition input screen 120. In the scoring data 18 described above with reference to FIG. 7, when the link type is “revise”, the user who created the original article has “3 points”, and the user who created the previous article has “7 points”. there were. On the other hand, as shown in FIG. 15A, in the scoring data 18b that emphasizes “revise”, when the link type is “revise”, the scoring of the user who created the original article is “5 points”. The score of the user who created the previous article is “10 points”, which is higher than the score data 18 shown in FIG. Therefore, the score of the user who created the article whose link type is “revise” in the score calculation unit 14 is higher than that calculated based on the score data 18 shown in FIG. 7. The calculation is based on the data 18b.

  The scoring data 18c of FIG. 15B is used when “sum” is emphasized, and is selected by the radio button 122b of the search condition input screen 120. In the scoring data 18 described above with reference to FIG. 7, when the link type is “sum”, the scoring of the user who created the original article is “3 points”, and the scoring of the user who created the previous article is “6 points”. there were. On the other hand, as shown in FIG. 15B, in the scoring data 18b that emphasizes “sum”, when the link type is “sum”, the scoring of the user who created the original article is “5 points”. The score of the user who created the previous article is “9 points”, which is higher than the score data 18 shown in FIG. Therefore, the score of the user who has created the article whose link type is “sum” in the score calculation unit 14 is higher than that calculated based on the score data 18 shown in FIG. The calculation is based on the data 18c.

  The scoring data 18d in FIG. 15C is used when “consult” is emphasized, and is selected by the radio button 122c on the search condition input screen 120. In the scoring data 18 described above with reference to FIG. 7, when the link type is “consult”, the user who created the original article has “3 points”, and the user who created the previous article has “0 points”. there were. On the other hand, in the scoring data 18d shown in FIG. 15C, when the link type is “consult”, the scoring of the user who created the original article is “5 points” and the user who created the previous article The score is “8 points”, which is higher than the score data 18 shown in FIG. Therefore, the score of the user who created the article whose link type is “consult” in the score calculation unit 14 is higher than that calculated based on the score data 18 shown in FIG. The calculation is based on the data 18d.

<Modification of information search>
In the above-described embodiment, the score calculation unit 4d is described as a method of adding a fixed score specified by the score data 18 to the searched article. On the other hand, the score calculation unit 14 may add the points in consideration of the number of articles including keywords when the scores are totaled. Specifically, the score calculation unit 14 may calculate the score by taking the number of keywords included in the searched article as a score and taking this score into account. As an example, a method of calculating a user's score by multiplying a value (x × 0.1 + 1) obtained by multiplying the number of keywords x included in an article by 0.1 and adding 1 (x × 0.1 + 1) as a score for each score is considered. It is done.

  For example, in the case of the example described above with reference to FIG. 10, “16 points” for article A, “0 points” for article B, and “3 points” for article C are added to the score of user a. Assume that the number of keywords in article A is “4” and the number of keywords included in article C is “2”. When the score is taken into account by the method described above, the score added to the article A is 1.4 and the score added to the article C is 1.3 when the score of the user a is calculated. Therefore, the calculation of the score by adding the score to the normal score is obtained by 16 points × 1.4 + 3 points × 1.2, and as a result, the score of the user a is 27 points.

  In addition, the score calculation unit 14 may use TF / IDF using natural language processing used in the keyword extraction unit 12.

  As described above, the embodiments of the present invention have been described. However, it should not be understood that the descriptions and drawings constituting a part of this disclosure limit the present invention. From this disclosure, various alternative embodiments, examples, and operational techniques will be apparent to those skilled in the art.

  It goes without saying that the present invention includes various embodiments not described herein. Therefore, the technical scope of the present invention is defined only by the matters described in the above description and the invention specific matters according to the obvious claims.

It is a figure explaining the structure of the information search system which concerns on the best embodiment of this invention. It is a figure explaining the article which is the information utilized in the information search system which concerns on the best embodiment of this invention. It is a figure explaining the link of the article utilized in the information search system which concerns on the best embodiment of this invention. It is a figure explaining the link classification of the article utilized in the information search system which concerns on the best embodiment of this invention. It is an example of the article data utilized in the information search system which concerns on the best embodiment of this invention. It is an example of the link data utilized in the information search system which concerns on the best embodiment of this invention. It is an example of the distribution data utilized in the information search system which concerns on the best embodiment of this invention. It is a flowchart explaining the information search method which concerns on the best embodiment of this invention. It is a figure explaining the article linked with the article searched in the information search system which concerns on the best embodiment of this invention. It is an example of the score data memorize | stored in the information search system which concerns on best embodiment of this invention. It is an example of the search result screen displayed in the information search system which concerns on the best embodiment of this invention. It is an example of the search condition input screen displayed in the information search system which concerns on the modification of the best embodiment of this invention. It is an example of the stipulation data adjusted with the search condition input screen of FIG. It is another example of the search condition input screen displayed in the information search system which concerns on the modification of the best embodiment of this invention. It is an example of the stipulation data utilized by the search condition input screen of FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Information retrieval system 10 ... Information retrieval apparatus 11 ... Input / output interface 12 ... Keyword extraction part 13 ... Information retrieval part 14a ... Link extraction part 14b ... Creator extraction part 14c ... Scoring extraction part 14d ... Score calculation part 15 ... Ranking determination Unit 16 ... Transmission processing unit 17 ... Scoring setting unit 18, 18 a to 18 d ... Scoring data 19 ... Scoring data 20 ... Information storage device 21 ... Article data 22 ... Link data 30 ... User terminal 40 ... Network 100 ... Search result screen 110, 120 ... Search condition input screen 111, 121 ... Search condition input part 112a-112e ... Scroll button 122a-122c ... Radio button 113, 123 ... Search button

Claims (21)

An information search system in a user terminal that transmits search conditions and an information search apparatus that searches for information that matches the search conditions,
The information search device includes:
A storage device for storing article data having a record including an identifier of an article, an identifier of a user who created the article, and a body of the article;
A storage device for storing link data having a record in which identifiers of a plurality of related articles and a link type for identifying the related type are associated;
A storage device for storing scoring data having a record for associating the link type and the scoring point specified by the link type;
When a search condition including a keyword is input, an information search unit that searches an article including the keyword as a target article from the article data;
A link extraction unit that extracts a record including the identifier of the target article from the link data, and determines an identifier of the article associated with the identifier of the target article as an identifier of the link article;
A score extracting unit that extracts a score associated with the link type included in the record extracted from the link data from the score data;
A creator extraction unit that extracts a user identifier associated with the identifier of the target article from the article data, and extracts a user identifier associated with the identifier of the link article;
A score calculation unit that aggregates the score extracted by the score extraction unit for each identifier of the user;
An information retrieval system comprising: For an original article that is an arbitrary article, an article generated in relation to the original article is a previous article,
The link data associates an identifier of an original article and an identifier of a previous article generated in association with the original article as identifiers of the related articles.
The scoring data associates the scoring with the type of link, the type of the original article or the type of the previous article,
The score distribution unit extracts the score associated with the link type and the original article type, and the score associated with the link type and the previous article type. Information retrieval system described.   When the identifier of the target article or the identifier of the user associated with the identifier of the linked article is extracted, the score calculation unit generates score data having a record that associates the identifier of the user with the score of the initial value, The information search system according to claim 1, wherein the score data is updated by adding the score extracted by the score extraction unit. The information search device includes:
The information search system according to claim 1, further comprising a keyword extraction unit that extracts a keyword from the search condition that is a sentence including the keyword. The information search device includes:
The system further comprises a score setting unit that changes a score of the score data according to the change ratio when a score change request including a score change rate of the score data is input together with the search condition. Information retrieval system. The storage device stores a plurality of scoring data generated for each scoring policy identified by a scoring data identifier,
The information search device includes:
The information search system according to claim 1, further comprising: a score setting unit that extracts score data identified by the identifier of the score data from the storage device when a score data identifier is input together with the search condition. .   The information search system according to claim 1, wherein the score calculation unit adds points in consideration of the number of articles including the keyword when counting the scores. An information search method used in an information search device for searching for information that meets a search condition,
Storing article data having a record including an identifier of an article, an identifier of a user who created the article, and a body of the article in a storage device;
Storing link data having a record in which a plurality of related article identifiers and a link type for identifying the related type are associated with each other in a storage device;
The point data having a record that associates the link type and the point specified by the link type is stored in a storage device;
When a search condition including a keyword is input, an article including the keyword is searched as a target article from the article data; and
A record including the identifier of the target article is extracted from the link data, and the identifier of the article associated with the identifier of the target article is determined as the identifier of the linked article;
A step of extracting a score associated with the link type included in the record extracted from the link data from the score data;
A user identifier associated with the identifier of the target article is extracted from the article data, a user identifier associated with the link article is extracted, and the score extracted from the score data is extracted for each identifier of the user. Steps to be aggregated,
A method for retrieving information, comprising: For an original article that is an arbitrary article, an article generated in relation to the original article is a previous article,
The link data associates an identifier of an original article and an identifier of a previous article generated in association with the original article as identifiers of the related articles.
In the scoring data, the scoring is associated with the type of link, the type of the original article or the type of the previous article,
In the step of extracting the score from the score data, the score specified by the link type and the original article type, and the score specified by the link type and the previous article type are extracted. 9. The information search method according to claim 8, wherein the information search method is performed.   When the identifier of the user who generated the target article or the linked article is extracted, the score data is generated as data having a record that associates the identifier of the user and the score of the initial value, and the scoring data is included in the score The information retrieval method according to claim 8, wherein the points extracted from the list are added and updated.   9. The information search method according to claim 8, further comprising a step of extracting a keyword from the search condition that is a sentence including the keyword. Along with the search condition, when a point change request including a point change ratio of the point data is input, the point of the point data is changed according to the change ratio;
The information search method according to claim 8, further comprising: In the step of storing the scoring data in the storage device, a plurality of scoring data generated for each scoring policy identified by the scoring data identifier is stored,
When the identifier of the scoring data is input together with the search condition, the scoring data identified by the identifier of the scoring data is extracted from the storage device;
The information search method according to claim 8, further comprising:   9. The information search method according to claim 8, wherein when the scores are totaled, points are added in consideration of the number of articles including the keyword. An information search program used in an information search device for searching for information that meets search conditions,
Storing article data having a record including an identifier of an article, an identifier of a user who created the article, and a body of the article in a storage device;
Storing link data having a record in which a plurality of related article identifiers and a link type for identifying the related type are associated with each other in a storage device;
Storing score data having a record associating the link type and the score specified by the link type in a storage device;
When a search condition including a keyword is input, a step of searching an article including the keyword as a target article from the article data;
Extracting a record including the identifier of the target article from the link data, and determining an identifier of the article associated with the identifier of the target article as an identifier of the linked article;
Extracting a score associated with the link type including the record extracted from the link data from the score data;
The identifier of the user associated with the identifier of the target article is extracted from the article data, the identifier of the user associated with the identifier of the linked article is extracted, and the score extracted from the score data is extracted. Tally for each user identifier;
An information search program for causing a computer to execute the above. For an original article that is an arbitrary article, an article generated in relation to the original article is a previous article,
The link data associates an identifier of an original article and an identifier of a previous article generated in association with the original article as identifiers of the related articles.
The scoring data associates the scoring with the type of link, the type of the original article or the type of the previous article,
In the step of extracting the scoring from the scoring data, the scoring associated with the link type and the original article type, and the scoring associated with the link type and the preceding article type are extracted. 16. The information search program according to claim 15, wherein   The score data, when an identifier of a user associated with the identifier of the target article or the identifier of the link article is extracted, generates score data having a record that associates the identifier of the user with a score of an initial value, The information search program according to claim 15, wherein the score is extracted by adding the score extracted by the score extraction unit.   16. The information search program according to claim 15, further comprising a step of extracting a keyword from the search condition that is a sentence including the keyword.   16. The information search according to claim 15, further comprising a step of changing a score of the score data according to the change ratio when a score change request including a score change rate of the score data is input together with the search condition. program. In the step of storing the score data in a storage device, a plurality of score data generated for each score policy identified by the identifier of the score data is stored,
16. The information search program according to claim 15, further comprising the step of extracting from the storage device the score data identified by the identifier of the score data when an identifier of the score data is input together with the search condition. 16. The information search program according to claim 15, wherein when the scores are totaled, points are added in consideration of the number of articles including the keyword.

Images