JP5083627B2 - Minority opinion extractor - Google Patents

Description

  The present invention relates to a technique for extracting a minor (minority) but valuable opinion from opinions received from users.

  In recent years, various services have been provided on the Internet. As one example, there is a Web site (hereinafter referred to as an evaluation site) that collects and publishes evaluations and opinions by general users regarding products, services, and other events. In such an evaluation site, when a product name or the like is input, evaluation information and opinion information about the product by a general user are displayed. For this reason, the use of the evaluation site has an advantage that it is possible to know the evaluation and opinion of the user who actually uses the product, not the information close to the seller of the product.

In a conventional evaluation site, it is possible to know how many good evaluations and bad evaluations exist for a certain event, and it is also possible to know individual opinions. However, with regard to individual opinions, it is only possible to browse the ones selected arbitrarily from the received opinions. For example, if there are many good evaluations about an event, what points should be noted There is a problem that it is difficult to know a specific opinion whether many people say it is good. On the other hand, even if there are minority opinions, there are also valuable opinions, but here there is a problem that it is difficult to extract because there are few opinions. Although there is a technique for collecting a minority opinion for a question such as a questionnaire (see Patent Document 1), it is not sufficient to extract a qualified minority opinion.
JP 2005-92616 A

  Therefore, an object of the present invention is to provide a minority opinion extraction apparatus that can extract a valuable opinion, although it is a minority, from a blog, a bulletin board SNS, or the like.

  In order to solve the above problems, in the present invention, the article data is stored in association with the writer and the category, the storage unit stores the keyword in association with the category, the keyword specifying unit for specifying the keyword, the evaluation target, An evaluation information specifying means for specifying a combination of evaluation expressions, and an object description for extracting an object explanatory word for explaining the evaluation object from article data stored in the storage means according to the combination of the specified evaluation object and the evaluation expression Word extraction means, appearance frequency counting means for counting the number of appearances of the extracted target explanation word, and target explanation sentence extraction for extracting a sentence having a target explanation word with a small number of appearances as a target explanation text as a result of the counting And a category to which the keyword specified by the keyword specifying means belongs. A good writer identifying means for identifying a good writer and a small number in which the writer of an article including the target explanatory text among the extracted target explanatory texts extracts the target explanatory text that is the expert writer as a minority opinion A minority opinion extraction apparatus having an opinion extraction means is provided.

  According to the present invention, a target description including a specified keyword, an evaluation target, and an evaluation expression and including a target explanatory word explaining the evaluation target is extracted, and a writer who is good at a category corresponding to the keyword is specified. Since the target explanation contained in the article written by a good writer is extracted as a minority opinion, it is possible to extract a precious opinion, although it is a minority, from the opinions received about a certain event There is an effect.

(1. Device configuration)
DESCRIPTION OF EXEMPLARY EMBODIMENTS Hereinafter, preferred embodiments of the invention will be described in detail with reference to the drawings. FIG. 1 is a configuration diagram of a minority opinion extraction apparatus according to the present invention. In FIG. 1, 10 is a storage means, 21 is a keyword designation means, 22 is an evaluation information designation means, 30 is a target explanation word extraction means, 40 is an appearance count counting means, 50 is a good writer identification means, and 60 is a subject explanation sentence extraction. , 70 is a minority opinion extraction means, 80 is an opinion presentation means, and 90 is an arithmetic processing means.

  The storage unit 10 stores the article data in the text format extracted from the blog, the bulletin board SNS, etc. in association with the article ID for specifying the article data and the writer ID for specifying the writer of the article. , Which stores the correspondence between article IDs and categories relating to articles prepared in advance, and the correspondence between keywords and categories, and is realized by a storage device such as a hard disk connected to a computer. The keyword specifying means 21 is used for specifying a keyword to be searched by the user. The evaluation information designating unit 22 designates a combination of an evaluation object and an evaluation expression. The evaluation expression is a word representing evaluation such as “desirable”, “cute”, “beautiful”, and the evaluation object is a word to be evaluated. The minority opinion extraction apparatus 100 uses the keyword input by the user using the keyword specifying means 21 and the evaluation target and evaluation expression input by the evaluation information specifying means 22 to obtain article data requested by the user from text-format article data. Extract. If the evaluation expression is "I want" and the evaluation target is "Information", "Information" is written as "I want" and the article data that contains keywords entered by the user is useful for minority opinions. Is extracted and presented to the user. The keyword specifying means 21 and the evaluation information specifying means 22 are realized by an instruction input device such as a mouse and a keyboard, an input device I / F, and an OS function. The arithmetic processing unit 90 is realized by a CPU and a memory of a computer. When the CPU reads a dedicated program into the memory and executes it, the target explanatory word extraction unit 30, the appearance count counting unit 40, the good writer specifying unit 50, the target Functions as the explanatory note extracting unit 60 and the minority opinion extracting unit 70 are realized.

  The target explanatory word extraction unit 30 extracts from the storage unit 10 the article data having the keyword specified by the user by the keyword specifying unit 21 and the evaluation target and evaluation expression specified by the user by the evaluation information specifying unit 22. To do. A target explanatory word that is a word explaining the specified evaluation target is extracted from the extracted article data. If the evaluation target is “information”, if the content indicated by “information” is “actor”, “actor” is extracted as the target explanatory word. The appearance count counting unit 40 counts how many times each target explanatory word extracted by the target explanatory word extraction unit 30 appears in all article data in the storage unit 10. The good writer specifying means 50 refers to the keyword-category correspondence table stored in the storage means 10 with the keyword specified by the keyword specifying means 21, specifies the category of the keyword, and the writer who is good at the specified category. The writer ID that often writes articles in the category is specified. The target explanatory sentence extraction unit 60 extracts, as the target explanatory sentence, a sentence including the target explanatory words whose appearance count is less than a predetermined number among the target explanatory words counted by the appearance count counting unit 40. The minority opinion extracting means 70 extracts the writer ID of the article data including the target explanatory sentence from the target explanatory sentences extracted by the target explanatory sentence extracting means 60, and the writer ID extracted by the specialty writer specifying means 50. Are extracted as a minority opinion.

  The opinion presentation unit 80 presents the target explanatory text extracted by the minority opinion extraction unit 70, and is realized by a display device such as a liquid crystal display, an output I / F, and an OS function.

  Here, the information memorize | stored in the memory | storage means 10 is demonstrated. FIG. 2 is a diagram illustrating an example of information stored in the storage unit 10. 2A is article information, and FIG. 2B is an article-category correspondence table. The category may be assigned to the article by the article writer or may be automatically classified after the article data is collected as text data. FIG. 2C shows a keyword-category correspondence table. The keyword-category correspondence table is prepared manually in advance. Alternatively, keywords and categories can be automatically associated from article data. As shown in FIG. 2A, as article information, an article ID for specifying article data and a writer ID for specifying the writer of the article are stored in association with the article data. The article data stored in the storage means 10 is actually input from a user who has accessed the evaluation site published on the Internet, and is used by the server computer that has established the evaluation site. The article ID is given to the article data transmitted from the user, and the ID of the transmission source user is registered in the storage means 10 as the writer ID. In addition, there are cases where the description content on a blog or bulletin board on the Web is acquired as article data and registered in the storage means 10. In addition, the questionnaire data collected from the user by the company may be digitized and registered in the storage means 10 as article data.

(2. Processing operation)
Next, the processing operation of the minority opinion extraction apparatus shown in FIG. 1 will be described. First, the user designates a keyword from the keyword designation means 21 of the minority opinion extraction apparatus 100 operating on the Web. Further, the user designates an evaluation object and an evaluation expression from the evaluation information designating unit 22. Then, the target explanatory word extraction unit 30 performs a full-text search of the article data in the storage unit 10 with the specified keyword, the specified evaluation target, and the evaluation expression, and finds all three of the specified keyword, evaluation target, and evaluation expression. Extract the article data you have.

  Subsequently, the target explanatory word extraction unit 30 performs syntax analysis of the extracted article. Then, as a result of the syntax analysis, a specific part of speech is identified as a target explanatory word that is an evaluation target explanatory word among words related to the evaluation target input by the user. The part of speech to be the target explanatory word is set in advance, and the target explanatory word extraction means 30 extracts the word of the set part of speech as the target explanatory word. The part of speech that should be the target explanatory word can be set as appropriate, but in the present embodiment, a general noun, a noun that becomes a sub-variant connection, an independent adjective, and the like are set. In the example illustrated in FIG. 2, “actor”, “recent status”, “fashion”, “private”, and “shooting progress” are extracted as target explanatory words.

  Next, the appearance count counting unit 40 counts the number of times the target explanatory word extracted by the target explanatory word extraction unit 30 appears as the target explanatory word in the storage unit 10 for each target explanatory word. Therefore, even if the word is the same as the target explanatory word, it is not counted unless it appears as the target explanatory word. In the example shown in FIG. 2, the target explanatory word “actor” is twice, the target explanatory word “recent” is once, the target explanatory word “fashion” is once, the target explanatory word “private” is once, and the target description is The word “shooting progress” is counted once. For counting the number of appearances, in addition to the method of simply counting in this way, a method of counting by a combination of target explanatory words may be used. For example, in the case of the article K0002 shown in FIG. 2, the evaluation target “information” appears twice, and the first “information” target explanatory words are “actor” and “fashion”. Since both are related to the same evaluation object “information”, the combination of “actor” and “fashion” is counted as one time.

  Next, the minority opinion extracting means 70 selects a predetermined number of target explanatory words with a small number of appearances counted by the appearance number counting means 40 from those with a small number of appearances. The target explanatory words with a small number of appearances often have the same number of appearances, and there are too many pieces of article data to display the article data including the target explanatory words to the user. Therefore, in order to extract a more useful minority opinion, a process for narrowing down article data is performed. Specifically, first, the target explanatory sentence extracting means 60 extracts a sentence including the target explanatory word selected as the target explanatory word as the target explanatory word from the article data as the target explanatory sentence. When a sentence is extracted from article data, the start position and end position of the sentence are specified by a known syntax analysis technique.

  On the other hand, the good writer specifying unit 50 uses the keyword specified by the keyword specifying unit 21 to refer to the keyword-category correspondence table in the storage unit 10 and specifies the category to which the specified keyword belongs. For example, it is assumed that “ΔΔCafe” is specified by the keyword specifying means 21. In this case, referring to the keyword-category correspondence table in FIG. 2C, “movie” is specified as the category. Subsequently, the good writer specifying unit 50 specifies a good writer for the specified category. Specifically, with reference to the article information shown in FIG. 2A, the article ID of an article written by each writer within a certain period is acquired. Subsequently, the article ID category is specified by referring to the article-category correspondence table shown in FIG.

  Next, for each writer, the ratio of the number of articles belonging to the target category to the number of articles written by the writer is calculated, and this is used as the strength. For example, it is assumed that there are 30 articles written by the writer “H0001” within a certain period, and of these, 15 articles belong to the category “movie”. In this case, the goodness of the writer “H0001” is calculated as 15/30 = 0.5. For each writer in which one or more articles are registered in the article information, the proficiency is calculated in the same manner, and the writer with the highest proficiency is identified as the writer who is good at the category. For example, it is assumed that the writer “H0002” is good at 0.67 and the writer “H0003” is good at 0.33. In this case, since the writer “H0002” has the highest strength, the writer “H0002” is identified as the good writer.

  When the good writer specifying unit 50 specifies a good writer for the specified keyword, the minority opinion extracting unit 70 extracts a small number of the target explanations extracted by the target explanatory note extracting unit 60 and written by the good writer. Selected as an opinion.

  When the minority opinion is extracted by the minority opinion extraction unit 70, the opinion presentation unit 80 creates a title including the target explanatory word and the evaluation target for each minority opinion. Specifically, it is created by extracting all character strings between the target explanatory word and the evaluation target from the article data. Then, the created title is arranged in a preset layout, and screen data is created and presented.

  The state of the screen presented at this time is shown in FIG. In FIG. 3, the keyword specified by the keyword specifying means 21, the evaluation target specified by the evaluation information specifying means 22 and the evaluation expression are displayed at the top of the screen. At the bottom of the screen, a list of “minority comments” is displayed. In the minority comments column, the titles of articles extracted as minority comments are listed.

  The minority opinion extraction apparatus according to the present invention can be used as a stand-alone type or a network type. For example, when questionnaire results are collected as article data and analyzed by a single computer to extract minority opinions, a stand-alone type is realized. Further, in order to allow general users to browse from their own terminals, a part of the keyword specifying means 21, the evaluation information specifying means 22, and the opinion presentation means 80 is configured by a network communication function, so that it is a network type. Realized.

It is a block diagram of the minority opinion extraction apparatus which concerns on this invention. 4 is a diagram illustrating an example of information stored in a storage unit 10. FIG. It is a figure which shows the mode of the screen shown by the presentation means.

Explanation of symbols

DESCRIPTION OF SYMBOLS 10 ... Memory | storage means 21 ... Keyword designation | designated means 22 ... Evaluation information designation | designated means 30 ... Target explanatory word extraction means 40 ... Appearance count counting means 50 ... Special note hand identification means 60 ... Object explanation sentence extraction means 70 ... minority opinion extraction means 80 ... opinion presentation means 90 ... arithmetic processing means 100 ... minority opinion extraction device

Claims (2)

Storing the article data in association with the writer and the category, and storing the keyword in association with the category;
A keyword specifying means for specifying a keyword,
An evaluation information specifying means for specifying a combination of an evaluation object and an evaluation expression;
A target explanatory word extracting means for extracting a target explanatory word for explaining the evaluation target from the article data stored in the storage means according to the combination of the designated evaluation target and the evaluation expression;
Appearance count counting means for counting the number of appearances of the extracted target explanatory words;
As a result of the counting, target explanation sentence extracting means for extracting a sentence having a target explanation word with a small number of appearances as a target explanation sentence;
Identifying a category to which the keyword designated by the keyword designation means belongs, and identifying a writer who is good at the identified category;
Of the extracted target explanatory texts, the writer of the article including the target explanatory text is a minority opinion extracting means for extracting the target explanatory text that is the expert writer as a minority opinion,
A minority opinion extraction device characterized by comprising:   A program for causing a computer to function as the minority opinion extracting device according to claim 1.

Images