JP2009086944A - Information processor and information processing program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an information processor which acquires evaluation information for a sentence instead of evaluation information such as reputation and criticism for a specific word (keyword), together with relationship with information (sentence) to be evaluated. <P>SOLUTION: The information processor 10 includes: a first information acquisition part 11 for acquiring first information including a sentence; a second information acquisition part 12 for acquiring second information to be evaluated including a sentence; an identification part 13 for identifying a quotation part and a related part related to the quotation part, in the first information; a similarity determination part 14 for determining the similarity between the quotation part and content of the second information; a related description amount determination part 15 for determining the description amount of the related part or the related description amount indicative of the amount of evaluation expression in the related part; and an output part 16 for outputting first information based on the similarity and a determination result of the related description amount. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to an information processing apparatus and an information processing program.

  There is a technique for extracting an evaluation of a target from a document (for example, see Patent Documents 1 and 2).

  Patent Document 1 describes a technique for extracting a pair of an impression expression word expressing an impression of an object and a noun phrase to which the impression expression word is connected from document data. In this technique, for example, when a lipstick is targeted, “good”-“feeling” is extracted as a pair of an impression expression word and a noun phrase.

  Patent Document 2 describes a technique for extracting reputation information about a search word from a document based on the appearance position of the search word and the appearance position of the evaluation expression in the document. In this technology, for example, reputation information “mobile gear is more useful” including a character string between a search word “mobile gear” and an evaluation expression “useful” is extracted. Patent Document 2 describes that priorities are assigned to the extracted pieces of reputation information in descending order of the probability that the pieces of reputation information are related to search terms.

  The website “blogWatcher” (http://blogwatcher.pi.titech.ac.jp/) of the Okumura Laboratory at the Institute of Precision Engineering, Tokyo Institute of Technology can be used to search for keywords such as product names from a large number of collected blogs. A service to extract reputation information is provided.

  In addition, the cyber agent website “Ameba Word-of-mouth Reputation Search” (http://buzz.ameba.jp/) is a graph showing the number of occurrences and likability of specific keywords that are on the blog. There is a service that can be displayed.

  Patent Document 3 describes a technique for finding a site containing a large amount of information (for example, a reputation for a company or a product) that matches the purpose of the survey from the Internet.

Japanese Patent No. 3787318 Japanese Patent No. 3820878 Japanese Patent Laid-Open No. 2004-280568

  The present invention relates to an information processing apparatus or information processing program that can obtain evaluation information for a sentence, not evaluation information such as reputation or criticism for a specific word (keyword), with the relevance to the evaluated information (sentence). The purpose is to provide.

  An information processing apparatus according to the present invention includes a first information acquisition unit that acquires first information including a sentence, a second information acquisition unit that acquires second information that includes a sentence and is an object of evaluation, A specifying means for specifying a cited part and a related part related to the cited part among the information of 1, a similarity determining means for determining a similarity of contents between the cited part and the second information, A related description amount determination means for determining a related description amount indicating a description amount of the related portion or an evaluation expression in the related portion; and first information based on the determination result of the similarity and the related description amount. Output means.

  In one aspect of the present invention, the similarity determination unit divides the second information into a plurality of parts, determines the similarity of the contents of the cited part and each part of the second information, Based on the determination result, it is determined which part of the second information the cited part is similar to, and the output means determines the part of the second information that the cited part is similar to Is output.

  In one aspect of the present invention, the related portion of the first information output by the output unit is displayed when the similarity is a predetermined value or more and the related description amount is a predetermined value or more. Display control means for displaying on the apparatus is further provided.

  In the aspect of the invention, the similarity determination unit divides the second information into a plurality of parts, and determines the similarity of contents between the cited part and each part of the second information. At the same time, based on the determination result, it is determined which part of the second information the cited part is similar to, and the display control means is similar to the cited part corresponding to the related part. The second information part is displayed in association with the second information part.

  An information processing program according to the present invention includes a step of acquiring, in a computer, first information including a sentence, a step of acquiring second information that includes a sentence and is an object of evaluation, and among the first information , Identifying a cited part and a related part related to the cited part, determining a similarity between contents of the cited part and the second information, a description amount of the related part or the related A step of determining a related description amount indicating the amount of evaluation expression in the portion; and a step of outputting first information based on the determination result of the similarity and the related description amount.

  In one aspect of the present invention, in the step of determining the similarity, the second information is divided into a plurality of parts, and the similarity of contents between the cited part and each part of the second information is determined. In addition, based on the determination result, it is determined which part of the second information the cited part is similar to, and in the outputting step, which part of the second information is the similar to the cited part The judgment result of is output.

  In one aspect of the present invention, the related portion of the first information output in the outputting step is determined when the similarity is a predetermined value or more and the related description amount is a predetermined value or more. It further has a step of displaying on the display device.

  In the aspect of the present invention, in the step of determining the similarity, the second information is divided into a plurality of parts, and the similarity of contents between the cited part and each part of the second information is determined. And determining, based on the determination result, which part of the second information the reference part is similar to, and in the step of displaying the reference part, the reference part corresponding to the relevant part is The second information is displayed in association with the similar second information portion.

  According to the first aspect of the present invention, it is possible to obtain evaluation information for a sentence to be evaluated in relation to information to be evaluated (sentence).

  According to invention of Claim 2, the evaluation information with respect to the part of the information evaluated can be obtained.

  According to invention of Claim 3, evaluation information can be displayed in relation to the information evaluated.

  According to invention of Claim 4, evaluation information can be displayed in relation to the part of the information evaluated.

  According to the fifth aspect of the present invention, it is possible to obtain evaluation information for a sentence to be evaluated in relation to information (sentence) to be evaluated.

  According to the sixth aspect of the present invention, it is possible to obtain evaluation information for a portion of information to be evaluated.

  According to invention of Claim 7, evaluation information can be displayed in relation to the information evaluated.

  According to invention of Claim 8, evaluation information can be displayed in relation to the part of the information evaluated.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a block diagram showing an example of the configuration of the information processing apparatus 10 according to the present embodiment. In one aspect, the information processing apparatus 10 is realized by cooperation of hardware resources and software, and is, for example, a computer. Specifically, the function of the information processing apparatus 10 is realized by reading an information processing program recorded on a recording medium into a main memory and executing it by a CPU (Central Processing Unit). The information processing program can be provided by being recorded on a computer-readable recording medium such as a CD-ROM, or can be provided by communication as a data signal. However, the function of the information processing apparatus 10 may be realized only by hardware. Further, the information processing apparatus 10 may be physically realized by one apparatus or may be realized by a plurality of apparatuses. In a specific aspect, the information processing apparatus 10 is a server apparatus on the Internet.

  In FIG. 1, the information processing apparatus 10 includes a first information acquisition unit 11, a second information acquisition unit 12, a specification unit 13, a similarity determination unit 14, a related description amount determination unit 15, and an output unit 16.

  The first information acquisition unit 11 acquires first information including a sentence.

  Here, the first information is specifically document information including sentences, and is an article collected from a website such as a blog.

  The first information acquisition unit 11 acquires first information from, for example, a storage device such as a hard disk device of the information processing device 10 or a device on the Internet.

  The first information acquisition unit 11 may acquire the first information from a predetermined storage area, or the information specified from outside (other devices, other program modules, users, etc.) It may be acquired as information.

  In one aspect, the 1st information acquisition part 11 acquires the 1st information group containing a plurality of 1st information.

  The second information acquisition unit 12 acquires second information that includes text and is to be evaluated.

  Here, the second information is specifically document information including a sentence, and more specifically is a natural sentence text in which fact information such as a news article or a press release is described.

  In one mode, the 2nd information acquisition part 12 receives the information specified by the user as the 2nd information from the client device via the Internet. Examples of the information specified by the user include text input from the user to the client device via an input device such as a keyboard, a document file in the client device specified by the user, and a paper document provided by the user. Text obtained by scanning and character recognition. At this time, the user may use a news article or a press release article on the Internet as the second information as it is.

  The second information acquisition unit 12 may acquire the second information in a mode other than the above. For example, the second information acquisition unit 12 may accept input of text from the user via the input device of the information processing device 10, and information based on an instruction from the user (or other device or program module) A document file in the processing apparatus 10 or a document file on the Internet may be acquired.

  The specifying unit 13 specifies a cited part and a related part related to the cited part in the first information acquired by the first information acquiring unit 11.

  In one aspect, the specifying unit 13 specifies a cited part in the first information, and specifies a related part based on the specific result.

  Moreover, in one aspect, the specifying unit 13 specifies a cited part and a related part for each first information included in the first information group acquired by the first information acquiring unit 11.

  The cited part is a description part in which fact information such as a news article or a press release is cited and described.

  The specifying unit 13 may specify one or more cited parts from the first information, and may specify a related part related to each specified cited part.

  Examples of the method for specifying the cited part include the following methods (a1) to (a3). In addition, the following method may be used independently and may be used in combination.

(A1) The citation part is specified based on the additional information (for example, tag) representing the citation added to the text in the first information. For example, when the first information is an HTML (HyperText Markup Language) document, the text enclosed by the <BLOCKQUEOT> tag is specified as the cited part.
(A2) The quoted part is specified based on a symbol (for example, parentheses) surrounding the specific text included in the text in the first information. For example, if the start parenthesis (for example, ““ ”) and the end parenthesis (for example,“ ””) in the text are extracted, and the number of characters in the text between the parentheses is a predetermined number or more (for example, 20 characters or more), The text between the brackets is identified as the quoted part.
(A3) The text in the first information in which a link to other information is set is specified as the cited part. For example, in an HTML document, when an entire sentence or entire paragraph is anchor text, there is a high possibility that it is a link to a news article or the like of another site, so that the anchor text is specified as a citation part.

  As a method of specifying the related part, there is a method of specifying the related part based on the composition of the sentence. For example, related parts using the fact that “in blogs, etc., are cited articles describing the original fact information, and then many sentence structures are described that describe opinions, evaluations, etc.” Is identified. Specifically, the following methods (b1) to (b3) are exemplified as a method for specifying the related portion. In addition, the following method may be used independently and may be used in combination.

(B1) The text from immediately after the citation part to the next citation part is specified as the related part related to the former citation part.
(B2) When the text between the parentheses is specified as the cited part, the part other than the cited part is specified as the related part in the paragraph including the cited part.
(B3) When there is only one citation part in the first information, all texts in the first information excluding the citation part are specified as related parts.

  However, the related part may be specified based on other than the composition of the sentence. For example, in the first information, a portion having a high degree of content similarity (for example, a similarity based on a vector space method) with the cited portion may be specified as a related portion.

  The similarity determination unit 14 determines the similarity of the content between the quoted portion specified by the specifying unit 13 and the second information acquired by the second information acquisition unit 12.

  In one aspect, the similarity determination unit 14 determines, for each of one or more cited parts in the first information specified by the specifying unit 13, the similarity between the contents of the cited part and the second information. judge.

  Moreover, in one aspect, the similarity determination part 14 determines the similarity of the content of each quotation part and 2nd information about each of the 1st information contained in a 1st information group.

  However, in a mode in which the similarity determination by the similarity determination unit 14 is performed after the determination of the related description amount by the related description amount determination unit 15 described later, the similarity determination unit 14 The similarity may be determined only for a cited portion corresponding to a related portion (for example, a related portion determined to have a description) having a related description amount equal to or greater than a predetermined value.

  The similarity may be expressed in two stages of whether or not they are similar, or may be expressed in three or more stages.

  The similarity determination unit 14 extracts, for example, important words from each of the second information and the quoted portion, and calculates the similarity between the two using a vector space method or the like. Here, the important word is specifically an important word for expressing the fact, and is, for example, an independent word, a proper expression (proprietary noun, date expression, price expression, etc.), a technical term, or the like. The calculated similarity may be the final content similarity, or the final content similarity may be obtained based on the calculated similarity. Note that the similarity determination unit 14 may obtain the similarity by a method different from the above.

  When the quoted portion is anchor text, the similarity between the anchor text and the second information or the title of the second information may be obtained, or the entire text of the linked text of the anchor text And the similarity with 2nd information may be calculated | required.

  The related description amount determination unit 15 determines the related description amount indicating the description amount of the related part specified by the specifying unit 13 or the amount of the evaluation expression in the related part.

  In one aspect, the related description amount determination unit 15 determines the related description amount for each of one or more related portions in the first information specified by the specifying unit 13.

  In one embodiment, the related description amount determination unit 15 determines the related description amount for each related portion of each first information included in the first information group. However, in the aspect in which the determination of the related description amount by the related description amount determination unit 15 is performed after the similarity determination by the similarity determination unit 14, the related description amount determination unit 15 includes the specified related portion. The related description amount may be determined only for a related part corresponding to a cited part (for example, a cited part determined to be similar) having a similarity equal to or greater than a predetermined value.

  The related description amount may be expressed in two stages of whether or not there is a description, or may be expressed in three or more stages.

  Examples of the method for determining the related description amount include the following methods (c1) to (c3).

(C1) A value corresponding to the description amount of the related part is obtained as the related description amount. For example, if the description amount is equal to or less than a predetermined amount (for example, one sentence or less, 10 characters or less, etc.), it is determined that there is no description (for example, related description amount = 0), and if not, there is a description (for example, related description) It is determined that amount = 1).
(C2) A value corresponding to the amount of the evaluation expression in the related part is obtained as the related description amount. Here, evaluation expressions are, for example, verbs that express opinions (“think”, “think”, “respect”, “expect”, etc.) and words that represent evaluation (“good”, “great”, “sorry”) Etc.).
(C3) The related description amount is obtained based on the description amount of the related portion and the amount of evaluation expression in the related portion. For example, the product or sum of the value according to the amount described above and the value according to the amount of the evaluation expression is calculated as the related description amount.

  By the way, the determination of the similarity by the similarity determination unit 14 and the determination of the related description amount by the related description amount determination unit 15 are performed on the first information in which the citation part is specified by the specifying unit 13. The first information for which the citation part is not specified by the specification unit 13 (hereinafter referred to as “first information not including the citation part”) may be excluded from the processing target or may be subjected to another process. May be.

  In one aspect, the information processing apparatus 10 includes the information evaluation unit 19, and thereby performs the following process on the first information that does not include the cited part.

  The information evaluation part 19 determines the value as evaluation information with respect to 2nd information of the said 1st information about each of the 1st information which does not include a quotation part. Evaluation information is information including an opinion, evaluation, reputation, or comment on fact information.

  Specifically, the information evaluation unit 19 determines the similarity between the contents of the first information and the second information for each of the first information that does not include the quoted portion. The determination method here may be the same as the determination method in the similarity determination unit 14, or may be another method.

  Further, the information evaluation unit 19 may determine the amount or content of the predetermined evaluation expression in the first information for each of the first information not including the quoted portion.

  The information evaluation unit 19 may use the determined similarity as the value as the evaluation information of the first information, or the first information based on the determined similarity and the amount and content of the evaluation expression. You may ask for value as evaluation information.

  The output unit 16 outputs first information based on the determination result of the similarity determination unit 14 and the determination result of the related description amount determination unit 15, that is, the determination result of the similarity and the related description amount.

  In one aspect, the information processing apparatus 10 includes the display control unit 17, the output unit 16 outputs the first information to the display control unit 17, and the display control unit 17 outputs the first information to the display device (not illustrated). Display information.

  In one aspect of the one aspect, the display control unit 17 has a related part of the first information output by the output unit 16 having a similarity of a predetermined value or more and a related description amount of a predetermined value or more. Display on the display device.

  However, the output unit 16 may output and store the first information to the storage device, or may output and print the first information to the printing device.

  The output unit 16 may execute the following processing in addition to the above processing, instead of the above processing, or as one aspect of the above processing.

  For example, the output unit 16 may output the determination result of the similarity and the related description amount. In this case, the output unit 16 may output the determination result to the storage device, store it, output it to the display device, display it on the screen, or output it to the printing device for printing. May be.

  The output unit 16 may output the similarity and the related description amount in association with each set of one or more cited parts and related parts.

  Further, the output unit 16 may output the specifying result of the specifying unit 13. For example, the output unit 16 may output the set of the specified citation part and the related part in association with the similarity and the related description amount corresponding to the set.

  The output unit 16 may obtain an evaluation value representing the value of the related part as evaluation information for the second information based on the similarity and the related description amount, and output the evaluation value as a determination result. . Here, the evaluation value is set as follows, for example. It is considered that the higher the similarity between the quoted part and the second information, the higher the possibility that the related part corresponding to the quoted part is the evaluation information for the second information. Therefore, the evaluation value of the related part is set to be higher as the similarity of the cited part corresponding to the related part is higher. In addition, when the quoted part and the second information are somewhat similar, the more the related description amount determined for the relevant part corresponding to the quoted part, the more the value of the relevant part as evaluation information for the second information Is considered high. Therefore, the evaluation value of the related part is set so as to increase as the related description amount of the related part increases. For example, the evaluation value is a product of the similarity and the related description amount.

  In addition, the output unit 16 includes, based on the determination result of the similarity determination unit 14 and the determination result of the related description amount determination unit 15, that is, the determination result of the similarity and the related description amount, Display information that preferentially displays related parts in which the degree of similarity of the corresponding cited parts is greater than or equal to a predetermined value and the related description amount is greater than or equal to a predetermined value may be output.

  Here, the output unit 16 may output and display the display information on a display device, or may output and display the information on a printing device, or may output and store the information on a storage device. Good.

  Further, the output unit 16 evaluates the second related information of the related part based on the similarity of the cited part corresponding to the related part and the related description amount of the related part for each specified related part. A related partial evaluation value representing the value as the above may be obtained, and display information may be created based on the related partial evaluation value. Here, the related portion evaluation value is set as follows, for example. It is considered that the higher the similarity between the quoted part and the second information, the higher the possibility that the related part corresponding to the quoted part is the evaluation information for the second information. Therefore, the related part evaluation value of the related part is set to be higher as the similarity of the cited part corresponding to the related part is higher. In addition, when the quoted part and the second information are somewhat similar, the more the related description amount determined for the relevant part corresponding to the quoted part, the more the value of the relevant part as evaluation information for the second information Is considered high. Therefore, the related part evaluation value of the related part is set so as to increase as the related description amount of the related part increases. For example, the related part evaluation value is a product of the similarity and the related description amount.

  The output unit 16 may create the display information so that the related portions are displayed in descending order of the related portion evaluation value. The output unit 16 may select a predetermined number of related parts in descending order of the related part evaluation value, and create display information so that the selected related parts are displayed. Further, the output unit 16 may select a related part having the related part evaluation value equal to or greater than a predetermined value, and create display information so that the selected related part is displayed.

  The display information may display the entire first information including a related portion to be displayed.

  The output unit 16 determines, for each first information, based on the similarity of the cited part in the first information and the related description amount of the related part in the first information. An evaluation value representing the value as evaluation information for the second information may be obtained, and display information may be created based on the evaluation value.

  The output unit 16 may create the display information so that the first information is displayed in descending order of the evaluation value, or the related parts in order from the information belonging to the first information having the highest evaluation value. The display information may be created so that is displayed. In addition, the output unit 16 selects a predetermined number of pieces of first information in descending order of the evaluation value, and creates display information so that related portions belonging to the selected first information are displayed. May be. Further, the output unit 16 selects the first information whose evaluation value is equal to or greater than the predetermined value, and creates display information so that the related part belonging to the selected first information is displayed. Good.

  Related parts other than the above-mentioned "related part in which the similarity of the corresponding cited part is equal to or greater than a predetermined value and the related description amount is equal to or greater than the predetermined value" may not be included in the display information, It may be included in the display information so as to be displayed in the subsequent stage of the related portion that is greater than or equal to the value.

  The output unit 16 may include the first information that does not include the quoted portion in the display information. For example, based on the value of the first information determined by the information evaluation unit 19, the output unit 16 selects the first information that does not include a predetermined number of citations in descending order of value, and the selected information is selected. The display information may be created so that the first information is displayed.

  Note that the output unit 16 may output various processing results such as a specific result, a determination result, a selection result, and a calculation result instead of outputting the display information. For example, the output unit 16 may associate the selected related part and the related part evaluation value of the related part or the rank of the related part evaluation value with each other and output them to the storage device. Further, the output unit 16 may output the selected first information and the evaluation value of the first information or the rank of the evaluation value to each other in association with each other. In addition, the output unit 16 outputs the selected first information and the value of the first information or the rank of the values to the storage device in association with each other for the first information that does not include the quoted portion. Also good.

  In the information processing apparatus 10 having the above configuration, in one aspect, the first information acquisition unit 11 acquires the first information group as follows.

  The first information acquisition unit 11 analyzes the second information acquired by the second information acquisition unit 12, and extracts an important word from the second information.

  Here, the important word is specifically an important word for expressing a fact, and is, for example, a proper expression or a technical term.

  The first information acquisition unit 11 may prioritize a specific part of speech such as a person's name or a place name in the extraction of an important word, may prioritize a frequently appearing part, or appear in the first sentence May be given priority.

  The first information acquisition unit 11 performs a search based on the extracted important word, and acquires information including the important word as first information.

  For example, the first information acquisition unit 11 retrieves and collects new articles from the Internet that include the extracted important words and within a predetermined period (such as within the last three days).

  In a more specific example, the first information acquisition unit 11 acquires an RSS feed from an RSS (Uniform Resource Locator) URL (RDF Site Summary) specified in advance or from an RSS URL collected by a crawler. Then, based on the creation date and time of new information described in the RSS feed, the URL of a newly arrived article (for example, registered within the last three days) is acquired. Then, the first information acquisition unit 11 accesses the acquired URL to acquire an article, and searches the acquired article including the extracted important word by AND search or OR search. To do. The first information acquisition unit 11 may rank the searched articles, select a predetermined number of articles in order from the top of the ranking, and use the predetermined number of articles as the first information group. In the ranking of the articles, for example, articles that include many important words are given priority, and articles that contain the same important words are given priority.

  The collection and search of new articles as described above can be realized by various methods such as using API (Application Program Interface) published by Technorati Japan Co., Ltd., for example.

  In addition, the 1st information acquisition part 11 may acquire a 1st information group by methods other than the above. For example, the first information acquisition unit 11 may acquire a first information group stored in advance in the storage device.

  FIG. 2 is a flowchart illustrating an example of an operation procedure of the information processing apparatus 10 according to the present embodiment. Hereinafter, the operation of the information processing apparatus 10 will be described with reference to FIG.

  In step S11, the information processing apparatus 10 acquires second information based on designation from the user.

  In step S12, the information processing apparatus 10 acquires first information from a predetermined storage device.

  Next, in step S13, the information processing apparatus 10 identifies one or more quoted parts and related parts related to each cited part in the acquired first information.

  Next, in step S14, the information processing apparatus 10 determines, for each of the identified citations, the similarity between the content of the citation and the acquired second information. Here, the degree of similarity is expressed in three stages of “0”, “1”, and “2”, and “0” indicates that they are not similar.

  Next, in step S15, the information processing apparatus 10 applies the citation portion determined to be similar to the second information among the identified related portions (that is, the citation portion having a similarity of “1” or “2”). For each corresponding related part, a description amount of the related part or a related description quantity indicating the amount of evaluation expression in the related part is determined. Here, the related description amount is expressed in four stages from “0” to “3”, and “0” indicates that there is no description.

  Next, in step S16, the information processing apparatus 10 causes the display device to display the first information based on the determination result of the similarity and the related description amount.

  In addition, when the cited part cannot be specified in step S13, the information processing apparatus 10 outputs information indicating that the cited part cannot be specified, for example, and ends the process.

  In one aspect of the present embodiment, the information processing apparatus 10 is as follows.

  In this aspect, the similarity determination unit 14 divides the second information into a plurality of parts, determines the similarity of the content between the cited part and each part of the second information, and based on the determination result, It is determined which part of the second information the cited part is similar to.

  The similarity determination unit 14 determines, for each of the specified one or more citation parts, the similarity between the citation part and each part of the second information, and based on the determination result, the citation part It may be determined which part of the second information the part is similar to.

  In addition, the similarity determination unit 14 determines the similarity of the content between the cited part and each part of the second information for each cited part of the first information of the first information group, and the determination Based on the result, it may be determined which part of the second information the cited part is similar to.

  The similarity determination unit 14 may divide the second information, for example, for each paragraph, or for each predetermined number of sentences. Moreover, the similarity determination unit 14 may divide the second information into a plurality of parts based on the description content of each area of the second information.

  The similarity determination unit 14 is, for example, a portion where the citation portion is similar to a portion having the maximum similarity with the citation portion based on the similarity in content between the citation portion and each portion of the second information. Is determined. In this case, when the maximum similarity is less than a predetermined value, it may be determined that the cited part is not similar to any part.

  In one aspect, the output unit 16 outputs a determination result as to which part of the second information the cited part is similar to.

  Further, in one aspect, the display control unit 17 displays the related part of the first information in association with the second information part having a similar citation part corresponding to the related part.

  Further, in one aspect, the output unit 16 displays the related part in association with the second information part similar to the cited part corresponding to the related part based on the determination result of the similarity determining unit 14. As described above, the display information is output.

  FIG. 3 is a flowchart showing another example of the operation procedure of the information processing apparatus 10 according to the present embodiment. Hereinafter, the operation of the information processing apparatus 10 will be described with reference to FIG.

(Step S31)
When the user wants to obtain evaluation information for a certain press release article, the user designates an HTML file in which the press release article is described to the client device. The client apparatus transmits the specified HTML file to the information processing apparatus 10 that is a server apparatus via the Internet.

  The information processing apparatus 10 receives the HTML file as second information (hereinafter referred to as “seed document”) from the client apparatus. The HTML file includes a title, a body HTML text, and a body text (a body HTML text obtained by removing a tag).

(Step S32)
Next, the information processing apparatus 10 analyzes the seed document by text analysis and extracts important words (here, a person name, an organization name, and a technical term) included in the first three sentences of the body text of the seed document.

(Step S33)
Next, the information processing apparatus 10 searches the Internet for new articles that contain the extracted important word and are within the last three days, and acquires them. Specifically, the information processing apparatus 10 acquires an RSS feed from a previously specified RSS URL of a search target, collects articles registered within the last three days based on the RSS feed, and collects the collected An article including an important word is searched from the article by OR search using the extracted important word as a search word. Then, the information processing apparatus 10 ranks the searched articles in the order including many search terms. At this time, articles with the same number of search terms are ranked in the order of creation date. Then, the information processing apparatus 10 acquires the top 50 articles (specifically, HTML files) in the ranking as search results. The information processing apparatus 10 holds the 50 acquired articles in association with the corresponding URLs. The article (HTML file) includes a title, a body HTML text, and a body text (a body HTML text obtained by removing a tag). The acquired article (HTML file) corresponds to the first information.

(Step S34)
Next, the information processing apparatus 10 analyzes the body HTML text to identify the cited part for each of the acquired articles, and identifies 50 articles included in the search result as the article including the cited part, and the cited part. It is divided into articles that do not contain.

  Here, the citation part is specified as follows. That is, if there is a part surrounded by <blockquote> tags in the body HTML text, the text of the part is specified as a cited part. When there is no <blockquote> tag, if there is a part in which all the paragraphs are anchor text, the anchor text is specified as a quoted part.

(Step S35)
Next, the information processing apparatus 10 specifies a related part corresponding to each cited part included in the article for each article including the cited part, and holds a pair of the cited part and the related part.

  Here, the information processing apparatus 10 sets the text between the citation part and the next citation part as a related part corresponding to the former citation part, and the text between the last citation part and the end of the article as the above. The related part corresponding to the last quoted part.

(Step S36)
Next, the information processing apparatus 10 determines the degree of similarity between each cited part included in the article and the seed document for each article including the cited part. Here, the information processing apparatus 10 determines the degree of similarity and determines which part of the seed document each citation part is similar to.

  Specifically, the information processing apparatus 10 performs determination as follows.

  The information processing apparatus 10 divides the seed document into paragraphs using the <p> tag of the body HTML text of the seed document as a clue, and extracts the text of each paragraph. In the example shown in FIG. 4, the seed document is divided into three parts, areas A, B, and C.

  Next, the information processing apparatus 10 extracts important words from the text of each area of the seed document, weights the important words using tf * IDF, and creates a word vector for each area. The important words extracted here may be the same as or different from the important words extracted in step S32. In addition, the information processing apparatus 10 similarly creates a word vector for each quoted portion of each article.

  Next, the information processing apparatus 10 uses the vector space method based on the word vector of each area of the seed document and the word vector of each citation part of each article, and each citation of each article and each article. Calculate the similarity between parts. Here, the similarity is normalized so that the similarity between the same documents is 100.

  Next, the information processing apparatus 10 determines which area of the seed document each citation part of each article is similar to based on the calculated similarity. For example, when determining about a certain quoted part, the information processing apparatus 10 judges that the part is similar to the area having the highest similarity with the cited part. However, if the maximum similarity is less than a predetermined value (40 in this case), it is determined that the cited part is not similar to the seed document.

  FIG. 4 illustrates one article, and the article includes cited portions Q1 and Q2 and related portions R1 and R2. When the similarity between each area A, B, C of the seed document and each cited part Q1, Q2 is as shown in FIG. 5, the cited part Q1 is similar to area A as shown in FIG. The quoted portion Q2 is determined to be similar to the area C.

  Next, the information processing apparatus 10 determines, based on the correspondence table between the similarity and the similarity score shown in FIG. 6, for each cited part of each article, from the highest similarity among the similarities determined for the cited part. Find similarity score. The similarity score “0” indicates that the cited part is not similar to the seed document.

  As illustrated in FIG. 7, the information processing apparatus 10 holds, for each cited part of each article, a similarity score of the cited part and information indicating an area where the cited part is similar.

(Step S37)
Next, the information processing apparatus 10 uses the amount of description of the related part for the related part corresponding to each cited part (cited part having a similarity score other than “0”) determined to be similar to the seed document in step S36. Based on this, the related description amount of the relevant part is determined.

  Specifically, the information processing apparatus 10 determines the related description of each related portion from the description amount (number of sentences) of each related portion based on the correspondence table of the related portion description amount and the related description amount score shown in FIG. Find the quantity score.

  As illustrated in FIG. 7, the information processing apparatus 10 holds the related description amount score of the related part corresponding to the cited part for each cited part of each article.

(Step S38)
Next, the information processing apparatus 10 calculates the similarity between the content of the article and the seed document for the article that does not include the cited part.

  Specifically, the information processing apparatus 10 extracts important words from the seed document and the article and calculates the degree of similarity between them as in step S36.

(Step S39)
Next, the information processing apparatus 10 scores the articles based on the determination results of steps S36 and S37 for the articles including the cited portion, and selects a predetermined number (here, seven) of articles based on the scores of the articles. To do. Specifically, a predetermined number of articles are selected in descending order of score.

  Here, the score SC of the article is calculated by the following formula (1).

In the above formula (1), N is the total number of citation parts included in the article, S _i is a similarity score corresponding to the i-th citation part of the article, and R _i is the i-th part of the article. It is a related description amount score corresponding to the cited part.

  As illustrated in FIG. 7, the information processing apparatus 10 holds the calculated score of each article.

  In addition, the information processing apparatus 10 holds the selected article. Specifically, the information processing apparatus 10 relates the selected article (HTML file) to the URL related to the article, the citation part and the related part related thereto, and each citation part is similar to which area of the seed document. Are stored together with the determination result, the score or the rank of the score.

(Step S40)
Next, the information processing apparatus 10 scores the articles based on the determination result of step S38 for the articles that do not include the quoted portion, and selects a predetermined number (three in this case) of articles based on the score of each article. . For example, the information processing apparatus 10 obtains a score corresponding to the degree of similarity obtained in step S38 for each article, and selects a predetermined number of articles in descending order of score.

  The information processing apparatus 10 holds the selected article. Specifically, the information processing apparatus 10 holds the selected article (HTML file) together with the URL, the score, or the rank of the score related to the article.

(Step S41)
Next, the information processing apparatus 10 creates display information in which the seed document and the article selected in steps S39 and S40 are laid out based on the various pieces of held information, and transmits the display information to the client apparatus via the Internet. And displayed on the screen of the client device.

  FIG. 9 shows an example of display. In FIG. 9, each area of the seed document is displayed separately, and below each area, related parts related to the cited part similar to the area are displayed in a list. The related parts corresponding to each area are displayed in order from the article belonging to the article with the highest score. Articles that do not include citations are displayed in descending order of score under the display of the seed document and related parts.

  In addition, this invention is not limited to the said embodiment, It can change variously within the range which does not deviate from the summary of this invention.

It is a block diagram which shows an example of a structure of the information processing apparatus which concerns on embodiment. It is a flowchart which shows an example of the operation | movement procedure of the information processing apparatus which concerns on embodiment. It is a flowchart which shows another example of the operation | movement procedure of the information processing apparatus which concerns on embodiment. It is a conceptual diagram which shows an example of a seed document and an article. It is a figure which shows an example of the determination result of the similarity of each area of a seed document, and each cited part. It is a figure which shows the correspondence table of similarity and a similarity score. It is a figure which shows an example of the content hold | maintained of information processing apparatus. It is a figure which shows the correspondence table of the description amount of a related part, and a related description amount score. It is a figure which shows an example of a display.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 10 Information processing apparatus, 11 1st information acquisition part, 12 2nd information acquisition part, 13 specific | specification part, 14 similarity determination part, 15 related description amount determination part, 16 output part, 17 display control part

Claims (8)

First information acquisition means for acquiring first information including a sentence;
Second information acquisition means for acquiring second information to be evaluated including sentences;
Of the first information, a specifying means for specifying a citation part and a related part related to the citation part;
Similarity determination means for determining the similarity of the content of the cited part and the second information;
A related description amount determination means for determining a related description amount indicating a description amount of the related portion or an evaluation expression amount in the related portion;
Output means for outputting first information based on the determination result of the similarity and the related description amount;
An information processing apparatus comprising: The information processing apparatus according to claim 1,
The similarity determination unit divides the second information into a plurality of parts, determines the similarity of contents between the cited part and each part of the second information, and based on the determination result, Determining which part of the second information the quoted part is similar to,
The output means outputs a determination result as to which part of the second information the cited part is similar;
An information processing apparatus characterized by that. The information processing apparatus according to claim 1,
Display control means for causing the display device to display the related portion of the first information output by the output means when the similarity is a predetermined value or more and the related description amount is a predetermined value or more; An information processing apparatus comprising: The information processing apparatus according to claim 3,
The similarity determination unit divides the second information into a plurality of parts, determines the similarity of contents between the cited part and each part of the second information, and based on the determination result, Determining which part of the second information the quoted part is similar to,
The display control means causes the related part to be displayed in association with a part of the second information in which the cited part corresponding to the related part is similar.
An information processing apparatus characterized by that. On the computer,
Obtaining first information including a sentence;
Obtaining second information to be evaluated, including sentences;
Identifying a citation part and a related part related to the citation part in the first information;
Determining a similarity between contents of the quoted portion and the second information;
Determining a related description amount indicating a description amount of the related portion or an evaluation expression amount in the related portion;
Outputting first information based on the determination result of the similarity and the related description amount;
An information processing program for executing An information processing program according to claim 5,
In the step of determining the degree of similarity, the second information is divided into a plurality of parts, the degree of similarity between the cited part and each part of the second information is determined, and based on the determination result , Determine which part of the second information the cited part is similar to,
In the outputting step, a determination result as to which part of the second information is similar to the cited part is output;
An information processing program characterized by that. An information processing program according to claim 5,
A step of causing the display unit to display the related portion of the first information output in the outputting step when the similarity is a predetermined value or more and the related description amount is a predetermined value or more. An information processing program characterized by that. An information processing program according to claim 7,
In the step of determining the degree of similarity, the second information is divided into a plurality of parts, the degree of similarity between the cited part and each part of the second information is determined, and based on the determination result , Determine which part of the second information the cited part is similar to,
In the displaying step, the related part is displayed in association with a part of the second information in which the cited part corresponding to the related part is similar.
An information processing program characterized by that.