JP2005032197A - Information analysis system and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an information analysis system and its method for enabling a user to simply obtain a desired analysis result of an actual status of word-of-mouth communication. <P>SOLUTION: An analysis server 103 collects sentences including word-of-mouth communication information submitted to a Web site by Internet users, and analyzes each submitted content (whether it is positive or negative in the meaning of the sentence content) related to a person having submitted the word-of-mouth communication and to words; and the actual statuses of word-of-mouth communication on the Internet are collected and analyzed by introducing a scheme for providing the result thereof as a report in the form of a list or a graph to a user computer 105. The analysis server 103 sets an analysis procedure for previously deriving potential needs and the like of claimers and consumers in the report; and a user browsing the received report by analysis report browsing software 112 of the user computer 105 can simply obtain the analysis result desired by the user by narrowing the collected results by setting conditions such as a keyword and a period according to the procedure. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an information analysis system and method. More specifically, by collecting word-of-mouth information that propagates on computer networks such as the Internet, providing the analysis results to the user as a report, and further narrowing down the results by conditions such as keywords and period, The present invention relates to an information analysis system, a server system thereof, a client system thereof, and an information analysis method related to a procedure that enables detection of a Kramer or a latent need.

Traditionally, companies' voices and word-of-mouth communication regarding their own products, their products, and products of competitors and rivals (communication of media, rumors and reputation, communication that is individually communicated from person to mouth; reference materials-new If you want to understand the Myungen Dictionary, Fifth Edition, Sanseido)
-Conducting a questionnaire-Requesting a marketing research company to conduct market research-Collecting voices received from customer consultation rooms, etc.

  On the other hand, as a method of collecting word-of-mouth information on the Internet, analyzing it, and providing the result as a report, there is an analysis method using Web clipping service and text mining. There is a word-of-mouth information analysis service using (registered trademark) (for example, see Non-Patent Document 1).

  Here, a specific site “Price. com ”(http://www.kakaku.com/), collecting text information, cutting out words by morphological analysis, and extracting words extracted by multivariate analysis or statistical methods, as well as quantitative indicators Analyzing by series. Moreover, the analysis result is stored in a CD (compact disk) -ROM (read only memory) and provided as a report to the customer.

  Trueteller V. 2 includes analysis functions such as word frequency ranking, drill-down analysis, main topic analysis, characteristic word analysis, mapping, detailed text display, group attribute analysis, time series monitoring, scoring, and reports. .

  On the other hand, in order to obtain word-of-mouth information about products, places, people, etc. that the general consumer is interested in, review posting sites (for example, “Price.com” and “At Cosmetics” (http: // (cosme.net/), etc.) or enter appropriate words on a search site such as Yahoo! (registered trademark) or Google (registered trademark).

  FIG. 7 is a system configuration diagram showing an example of a word-of-mouth information analysis system for realizing a conventional word-of-mouth analysis service. FIG. 8 is a signal sequence diagram corresponding to the word-of-mouth information analysis system of FIG.

  In FIG. 7, a speaker computer 701 that is a client system of a speaker who speaks word-of-mouth information, a Web server 702 that operates a Web site where word-of-mouth information is disclosed, and an analysis server 703 are connected to a computer network such as the Internet. 704 is connected. The speaker computer 701 is provided with a browser 706 capable of speaking word-of-mouth information. The Web server 702 operates Web sites such as news, personal HP (homepage), and BBS (bulletin board system), and includes BBS software 707 that receives word-of-mouth information. The analysis server 703 includes a sentence collection module 708, a word extraction module 709, a word analysis module 710, and an analysis report creation module 711.

  The user computer 705, which is a client system not connected to the computer network 704 of the user who wants to obtain word-of-mouth analysis information, is provided with analysis report browsing software 712 for referring to the analysis report. .

  In FIG. 8, the speaker uses the browser 706 of the speaker computer 701 to speak word-of-mouth information to the BBS software 707 in the Web server 702. The utterance is accumulated in the Web server 702 (step S800). The analysis server 703 accepts the word-of-mouth collection request from the user of the user computer 705 offline by mail or the like or temporarily online by the user computer 705 (step S802).

  Next, the analysis server 703 collects text sentences including word-of-mouth information from the BBS software 707 in the Web server 702 using the sentence collection module 708 (steps S804 to S806). The collected text sentences are transmitted from the sentence collection module 708 to the word extraction module 709, and are divided into word units such as nouns by a method such as morphological analysis by the word extraction module 709 (step S808).

  The extracted words are transmitted from the word extraction module 709 to the word analysis module 710, and the word analysis module 710 calculates the correlation value between the words according to the appearance frequency and the statistical method, for each word type (step). S810). The result analyzed in the word analysis module 710 is transmitted to the analysis report creation module 711, and the analysis report creation module 711 creates an analysis report in which the result is arranged in a table or graph format (step S812). The analysis report is stored in the CD-ROM and delivered to the user offline (step S814).

  The user who has received the CD-ROM browses the analysis report using the analysis report browsing software 712 of the user computer 705 (step S816). At this time, menus such as word frequency ranking and drill-down analysis are prepared in the analysis report, but for example, the user is a Kramer (for example, someone who continues to complain or complain about a specific company or product) In order to search for the potential needs of consumers, there is no analysis procedure for that purpose, so it is necessary for the user himself to analyze the results through trial and error from various angles.

Nomura Research Institute, “Word of mouth information analysis service”, [online], [searched July 4, 2003], Internet <URL: http://www.trueteller.net/kuchikomi/kdc/index.html>

  In the above-mentioned conventional technology, in the conventional questionnaire survey and market survey etc., the user's real intention is not necessarily heard, and since the hypothesis is set at the time of creating the questionnaire, new discoveries outside the hypothesis cannot be obtained, etc. There was a problem.

  On the other hand, in order to collect word-of-mouth information on the Internet, it is necessary to register keywords in advance for a company that provides word-of-mouth information analysis services. Since the analysis is performed and the result is provided to the customer by a CD-ROM or the like, there is a problem that information cannot be obtained immediately when the customer wants to know.

  In addition, when a general consumer tries to obtain information using a word-of-mouth posting site, it is not always the case that a word of interest is posted on the first visited site. Needed to walk through multiple word-of-mouth sites.

  And since there is no specific analysis procedure for obtaining information such as Kramer and the potential needs of consumers from collected word-of-mouth information, it was necessary for customers to analyze them from trial and error from various angles .

  The present invention has been made in view of such a problem, and an object of the present invention is to provide an information analysis system and method that can easily obtain the analysis result of the actual state of word of mouth desired by the user. It is in.

  In order to achieve such an object, in the information analysis system of the present invention, a first server system connected to a computer network accepts information spoken by a speaker and converts a text sentence including the accepted information. A client system that collects and analyzes the text sentence from one or more second server systems having sources published on the computer network, creates a report, and is connectable to the computer network. An information analysis system for displaying information based on the report, wherein the first server system includes at least a body including the information, an identifier of a speaker of the body, and a statement made by the speaker. The first identifier to be analyzed during the speech period of the body Based on the analysis means and the analysis result obtained by the first analysis means, at least the number of time-series utterances in the source that has been uttered is counted for each speaker, and the feature of the utterance content is specified. The client system includes: a second analysis unit; and a report creation unit that creates the report by graphing at least the analysis results obtained by the first analysis unit and the second analysis unit. , Based on the report, displaying a graph showing the number of utterances for each speaker in chronological order, period selection means for accepting selection of a utterance period with a large number of utterances from the displayed graph, and based on the report, A graph in which the speakers during the period selected by the period selection means are grouped according to the number of utterances of the speaker and the characteristics of the utterance contents is displayed. Group selection means for accepting selection of a specific group from the displayed graph, and based on the report, the sum total of the number of utterances of speakers in the specific group selected by the group selection means with respect to the total number of utterances A ratio selection means for displaying a graph showing a ratio and accepting a selection for the ratio of the specific group from the displayed graph, and the number of utterances of the specific group selected by the ratio selection means based on the report Based on the report, a source selection means for displaying a distribution, showing a graph showing the characteristics of the content of the utterance in time series for each of the said uttered sources, and receiving a selection of the specific source from the displayed graph, The number of utterances for each speaker in the source selected by the source selection means Information display means for displaying a graph showing the classification, wherein the report creation means displays the analysis result obtained by the first analysis means and the second analysis means by each means of the client system. The report is arranged in the form of a graph to be generated, and data setting for instructing the client system to sequentially execute each means of the client system is set in the report in advance. To do.

  Here, the feature of the statement content is a frequency indicating whether there are a lot of positive statements or a lot of negative statements, and the specific group has a large number of statements by the speaker and is negative as a feature of the statement content. The specific source is the source that has the most negative comments as a feature of the content of the speech, and whether the content of the speech is at least a positive or negative statement. It can be characterized by the distinction.

  With the above configuration, the Kramer can be detected.

  In order to achieve the above object, the information analysis system of the present invention is configured such that a first server system connected to a computer network accepts information spoken by a speaker and reads a text sentence including the accepted information. A client system that collects and analyzes the text sentence from one or more second server systems having sources published on the computer network, creates a report, and is connectable to the computer network. An information analysis system for displaying information based on the report, wherein the first server system includes at least a body including the information, an identifier of a speaker of the body, and a statement made by the speaker. The first identifier to be analyzed during the speech period of the body First, a word is extracted from the body obtained by the analyzing means and the first analyzing means, and at least the number of chronological statements in the source where the word is spoken is counted for each extracted word. 2 analysis means; and report creation means for creating the report by graphing at least the analysis results obtained by the first analysis means and the second analysis means, and the client system includes: Based on the report, a graph showing the number of utterances of a predetermined word for each source in time series is displayed, period selection means for receiving selection of a utterance period with a large number of utterances from the displayed graph, and the report On the basis of the statement including the predetermined word in the period selected by the period selection means, Displaying a graph in which the ratio of the number of utterances is grouped, a group selection unit that accepts a selection of a specific group from the displayed graph, and the specific group selected by the group selection unit based on the report Display a graph showing the distribution of the number of utterances for each of the said utterances in chronological order according to the feature of the utterance content, and source selection means for accepting the selection of a specific source from the displayed graph, and the report An information display means for displaying a graph showing the number of utterances for each word in the source selected by the source selection means for each utterance content of the body including the word, and the report creating means, Analysis results obtained by the first analysis means and the second analysis means are used as the client system. The report is arranged in the form of a graph displayed by each means of the above, and a data setting for instructing the client system to sequentially execute each means of the client system is set in the report in advance. It is characterized by leaving.

  Here, the statement content is either positive or negative, and the feature of the statement content is a frequency indicating whether there are many positive statements or many negative statements, and the specific content The source may be the source having the most positive comments as a characteristic of the content of the speech, and the content of the speech may be at least a distinction between a positive speech and a negative speech. .

  With the above configuration, it is possible to detect potential needs for a predetermined word (keyword).

  In order to achieve the above object, the information analysis method of the present invention is such that the first server system connected to the computer network accepts information spoken by a speaker and converts a text sentence including the accepted information. A client system that collects and analyzes the text sentence from one or more second server systems having sources published on the computer network, creates a report, and is connectable to the computer network. An information analysis method of an information analysis system for displaying information based on the report, wherein the first server system collects at least the text sentence including the information, an identifier of a speaker of the text, The identifier of the source spoken by the speaker, and the answer during the speech period of the text The first analysis step and the first server system, based on the analysis result obtained in the first analysis step, at least for each speaker, the number of utterances in time series in the source A second analysis step for identifying the feature of the statement content and the first server system graphing at least the analysis results obtained in the first analysis step and the second analysis step A report creation step for creating the report, and based on the report, the client system displays a graph showing the number of utterances for each speaker in chronological order, and makes a statement from the displayed graph. A period selection step for accepting selection of a speech period with a large number of cases, and the client system, based on the report, The group selection which displays the graph which grouped the speaker during the period selected in the recording period selection step according to the number of the speaker's utterances and the feature of the utterance content, and accepts the selection of a specific group from the displayed graph And a graph showing the ratio of the sum of the number of utterances of the speakers in the specific group selected in the group selection step based on the report to the total number of utterances, and the display A ratio selection step of accepting a selection for the ratio of the specific group from the generated graph, and the client system is remarked based on the report, the message number distribution of the specific group selected in the ratio selection step. In addition, according to the features of utterance contents in time series for each source A source selection step of receiving a selection of the specific source from the displayed graph, and for each speaker in the source selected in the source selection step based on the report An information display step for displaying a graph indicating the number of utterances by utterance content, and the server system uses the analysis results obtained in the first analysis step and the second analysis step in the report creation. In the report, a data setting for instructing the client system in order to sequentially execute each step of the client system is created in the form of a graph displayed in each step of the client system. It is characterized by being set to.

  In order to achieve the above object, the information analysis method of the present invention is such that the first server system connected to the computer network accepts information spoken by a speaker and converts a text sentence including the accepted information. A client system that collects and analyzes the text sentence from one or more second server systems having sources published on the computer network, creates a report, and is connectable to the computer network. An information analysis method of an information analysis system for displaying information based on the report, wherein the first server system collects at least the text sentence including the information, an identifier of a speaker of the text, The identifier of the source spoken by the speaker, and the answer during the speech period of the text The first analysis step, and the first server system extracts a word from the body obtained in the first analysis step, and at least the word is said for each extracted word A second analysis step for counting the number of messages in time series in the source; and the first server system graphs at least the analysis results obtained by the first analysis step and the second analysis step. A report creation step for creating the report, and based on the report, the client system displays a graph showing the number of utterances of a predetermined word for each source in time series, and the displayed A period selection step of accepting selection of an utterance period with a large number of utterances from the graph, and the client system includes the report And displaying a graph in which the ratio of the number of utterances is grouped according to the utterance content for the predetermined word for the body including the predetermined word during the period selected in the period selection step. A group selection step for accepting a selection of a specific group from the obtained graph, and the client system, based on the report, the message number distribution of the specific group selected in the group selection step for each of the said sources A source selection step of displaying a graph showing the utterance contents according to characteristics in time series and receiving a selection of a specific source from the displayed graph, and the source selection step based on the report by the client system Remarks for each word in the source selected in An information display step of displaying a graph showing the number of cases by content of remarks, and the server system, in the report creation, the analysis results obtained in the first analysis step and the second analysis step, The report is arranged in the form of a graph displayed at each step of the client system, and data settings for instructing the client system to sequentially execute the steps of the client system are set in the report in advance. It is characterized by keeping.

  In one embodiment of the present invention, a server system (analysis server) collects text including word-of-mouth information posted by an Internet user on a website, and posts related to the person or word that posted the review (in terms of comment content, specific (In terms of positive or negative), and introduces a system that provides the results to the client system (user computer) as a report such as a table or graph, and collects and analyzes the actual state of reviews on the Internet It is characterized by doing.

  Here, the server system sets in advance the analysis procedure that derives the potential needs of Kramer and consumers in the report, and the user who browses the received report with the browsing software of the client system follows the procedure such as keyword and period etc. The analysis result desired by the user can be easily obtained by setting conditions and narrowing down the collection results.

  With the above-described configuration of the present invention, the server system collects sentences for the bulletin board on the Internet, chat, and similar word-of-mouth services among users, and the analysis procedure preset by the user using the client system By searching the collection results according to the above, it becomes possible to search for the content related to the word of mouth and the word of the word of mouth about the word to be examined (specifically, positive or negative in terms of the content of the comment).

  As described above, according to the present invention, it is possible to obtain an analysis result of the actual state of the word of mouth desired by the user.

  Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In addition, the same code | symbol is attached | subjected to the location which has the same function in each drawing, and duplication of description is abbreviate | omitted.

  FIG. 1 is a system configuration diagram showing a word-of-mouth information analysis system for realizing a word-of-mouth analysis service according to an embodiment of the present invention. FIG. 2 is a signal sequence diagram corresponding to the word-of-mouth information analysis system of FIG.

  In FIG. 1, a server system that operates one or a plurality of speaker computers 101 which are client systems of speakers who speak word-of-mouth information, and an information source website where word-of-mouth information and the like are disclosed. One or a plurality of Web servers 102 and an analysis server 103 which is a server system for analysis are connected to a computer network 104 such as the Internet. The speaker computer 101 is provided with a browser 106 capable of speaking word-of-mouth information. The Web server 102 operates Web sites such as news, personal HP, and BBS, and includes BBS software 107 that accepts comments on word-of-mouth information. The analysis server 103 includes a sentence collection module 108, a word extraction module 109, and an analysis report creation module 111.

  A user computer 105, which is a client system of a user who wants to obtain word-of-mouth analysis information, can be connected to the computer network 104 by wire or wirelessly, and includes analysis report browsing software 112 for referring to the analysis report. Has been. Further, the analysis server 103 includes a sentence analysis module 201, a speaker analysis module 202, and a database 203 for storing collected word-of-mouth information.

  In each of the client systems and server systems described above, a CPU (central processing unit) (not shown) operates by reading a program loaded in a system memory (not shown). That is, in each client system and each server system, each CPU has a predetermined program (browser 106, BBS software 107, sentence collection module 108, word extraction module 109, analysis report creation module 111, sentence analysis module 201, speaker analysis). The module 202 operates according to instructions of the analysis report browsing software 112). Each of the modules included in the analysis server 103 is intended as a functional module constituting one software. As another form, each module or a combination of several modules may be independent software.

  The processing procedure of the word-of-mouth information analysis system in FIG. 2 shows the processing executed by each CPU and its linkage, and is executed by each CPU reading and executing a program loaded in the system memory. In order to send and receive data in the sequence of FIG. 2, a normal protocol such as TCP / IP (transmission control protocol / internet protocol), HTTP (hypertext transfer protocol), SMTP (simple mail transfer protocol), FTP (file transfer protocol) is used. do it.

  In FIG. 2, when a speaker utters word-of-mouth information using the browser 106 of the speaker computer 101, the CPU of the speaker computer 101 transmits the word-of-mouth information from the browser 106 to the BBS software 107 in the Web server 102. The word-of-mouth information is accumulated in a storage device (not shown) of the Web server 102 as a text sentence including the word-of-mouth information by the CPU of the Web server 102 (step S200). The CPU of the user computer 105 transmits the review collection request received from the user by the analysis report browsing software 112 to the analysis server 103, and the CPU of the analysis server 103 receives the review collection request (step S202).

Next, the CPU of the analysis server 103 collects text sentences including word-of-mouth information from the BBS software 107 in the Web server 102 using the sentence collection module 108 (steps S204 to S206). The CPU of the analysis server 103 transmits the collected text sentences from the sentence collection module 108 to the sentence analysis module 201 (step S208).
(1) Text including word-of-mouth information (speech)
(2) Information that identifies the speaker of the text (speaker name, etc.)
(3) Information for identifying the Web site that the speaker of the text said (Web site address, etc.)
(4) The speech period of the text (for example, the date and time, or the week of the latest 4 months, etc.)
Are stored in the database 203 as the analysis result 1 (step S210).

Further, the CPU of the analysis server 103 transmits the text (utterance) including the word-of-mouth information of (1) above from the sentence analysis module 201 to the word extraction module 109 (step S212), and divides it into words by morphological analysis, The word is extracted and returned to the sentence analysis module 201 (step S218). The CPU of the analysis server 103 uses the analysis result 1 by the sentence analysis module 201 for each word extracted.
・ Number of texts (utterances) containing the word
・ Number of people who made a statement ・ Time-series data (number of utterances in time series)
The actual condition of the word of mouth consisting of the above is measured and stored in the database 203 as the analysis result 3 (step S220).

The CPU of the analysis server 103
-Information about the speaker (information identifying the speaker in (2) above, and the text including the word of mouth information in (1) above)
Information identifying the Web site (3) related to the speaker. The message period (4) related to the speaker is transmitted from the sentence analysis module 201 to the speaker analysis module 202 (step S214). Based on these, the CPU of the analysis server 103 uses the speaker analysis module 202 to determine, for each speaker, word-of-mouth (corresponding to the number of comments) and the content of the comments (positive or negative). Analysis of the activity level (number of chronological utterances on the website) of the speaker who is speaking and the characteristics of the utterance content of the speaker (whether there are many positive or negative comments) Frequency) is specified and stored in the database 203 as analysis result 2 (step S216).

  The CPU of the analysis server 103 acquires the analysis results 1 to 3 stored in the database 203 by the analysis report creation module 111 (step S222), and uses them as a graph of each step in FIGS. Create analysis reports arranged in tables and graphs. The CPU of the analysis server 103 can transmit the analysis report to the analysis report browsing software 112 of the user computer 105 via the computer network 104 (step S224). As another form, after the analysis report is stored in a medium such as a CD-ROM, the analysis report may be sent offline to the user computer 105 and read by the analysis report browsing software 112.

  The CPU of the user computer 105 receives the analysis report from the analysis report creation module 111 using the analysis report browsing software 112, displays the analysis report on a display device (not shown), and performs browsing processing for the user. Here, the CPU of the analysis server 103 creates an analysis report for setting data for instructing an analysis procedure (to be described later with reference to FIGS. 3 to 6) of Kramer detection and consumer potential needs detection executed by the analysis report browsing software 112. It is set in the analysis report in advance by the module 111.

  More specifically, in this data setting, as shown in FIGS. 3 to 6 described later, an input event from the user to the analysis report browsing software 112 is detected, and each step is drawn and information is presented. This is data setting. If the analysis report browsing software 112 has a function of a general Internet browser, this data setting may be constructed by, for example, markup language HTML (hypertext markup language) or XML (eXtensible Markup Language), Alternatively, it may be constructed in a format unique to the analysis report browsing software 112.

  The user of the user computer 105 selects a procedure (described later with reference to FIGS. 3 to 6) based on the data setting that instructs the analysis procedure using the analysis report browsing software 112 when browsing the analysis report. It is possible to easily obtain a desired result by setting conditions such as a keyword and a period to narrow down the result.

  FIGS. 3 and 4 are examples of analysis procedures that are set in advance in the analysis report by the analysis report creation module 111 when the user of the user computer 105 detects Kramer in the embodiment of the present invention. . The CPU of the user computer 105 receives the analysis report from the analysis report creation module 111 using the analysis report browsing software 112, and executes this analysis procedure when performing browsing processing for the user. In the analysis procedure, first, in response to the reception of the analysis report by the analysis report browsing software 112, the “result display screen” and “condition setting” in <Step 1> shown in FIG. "Screen" is displayed and started. A description of each step is given below.

<Step 1 (Top page)>
When a user browses an analysis report using the analysis report browsing software 112, first, a graph 301 showing the number of messages for each website in time series (for example, weekly) is displayed on the “result display screen”. This represents the overall picture of the collection results. The user selects the “Kramer detection” menu 302 on the “condition setting screen”.

<Step 2 (Top page of Kramer search)>
On the “result display screen”, a graph 303 showing the number of utterances for each speaker in chronological order is displayed. The user selects an attention period (for example, period A indicated by reference numeral 304 in the figure) having a large number of cases on the “condition setting screen”.

<Step 3 (categorization of speakers)>
On the “result display screen”, all speakers who have spoken during the period selected in the previous step are grouped by the number of utterances and the characteristics of the utterances (whether there are many positive utterances or many negative utterances). A graph 305 categorized into D is displayed. Each speaker is displayed as one plot, and in which area of the graph 305 the plot is displayed, it can be known which category it belongs to. By indicating the group A indicated by reference numeral 306 in the “condition setting screen” to an area (group A in the first quadrant) having the highest number of statements and the most negative statements that are likely to belong to Kramer. select.

<Step 4 (confirm the percentage of target speech)>
On the “result display screen”, a graph 307 showing the ratio of the total number of utterances of the speakers of group A displayed in the first quadrant selected in the previous step to the total number of utterances is displayed. In this way, it is determined whether the number of statements that are highly likely to belong to Kramer is negligible, or is notable. By instructing the group A indicated by reference numeral 308 on the “condition setting screen”, the area “group A” is selected as the target area.

<Step 5 (Distribution of comments for each website)>
On the “Result Display Screen”, the distribution of the number of utterances of group A of utterances that are likely to belong to Kramer, and the characteristics of the utterance contents in chronological order for each website (there are many positive utterances or negative utterances) The graph 309 shown separately is displayed. On the “condition setting screen”, the Web site 310 with the most negative comments is selected. Here, the ellipses over a certain period in the graph 309 indicate that the larger the number, the larger the number of statements over that period, and the more dark the ellipse color, the more negative comments. .

<Step 6 (contents of speech for each speaker)>
On the “result display screen”, the number of utterances for each speaker on the Web site of interest selected in the previous step is displayed by utterance content (whether positive, negative, or otherwise moderate) Is displayed. The bar in the graph 311 indicates that the darker the color, the more negative the statement.

  The user performs the above-described series of operations, and when a speaker having a large number of negative statements is finally displayed on the “result display screen” in step 6, the speaker is said to be a Kramer. It leads to detection.

  FIGS. 5 and 6 are examples of analysis procedures set in advance in the analysis report by the analysis report creation module 111 when the user of the user computer 105 detects latent needs in the embodiment of the present invention. is there. The CPU of the user computer 105 receives the analysis report from the analysis report creation module 111 using the analysis report browsing software 112, and executes this analysis procedure when performing browsing processing for the user. In the analysis procedure, first, in response to the reception of the analysis report by the analysis report browsing software 112, the “result display screen” and “condition setting” in <Step 1> shown in FIG. "Screen" is displayed and started. A description of each step is given below.

<Step 1 (Top page)>
When a user browses a report using the analysis report browsing software 112, first, a graph 501 showing the number of messages for each website in time series (for example, weekly) is displayed on the “result display screen”. , Represents the overall picture of the collection results. The “latent need detection” menu 502 is selected on the “condition setting screen”, and the target keyword 503 is further input.

<Step 2 (Top page of Kramer search)>
The “result display screen” displays a graph 504 obtained by extracting the number of utterances including the keyword (for example, the personal computer X) from the graph 501 of the previous step. On the “condition setting screen”, an attention period with a large number of cases (for example, period A with reference numeral 505) is selected.

<Step 3 (Classification of positive / negative)>
“Result display screen” is a graph that categorizes the ratio of the number of utterances according to the utterance contents (“Positive” or “Negative”) for the utterances including the keyword (PC X) during the period selected in the previous step. 506 is displayed. The “positive” group 507 is selected on the “condition setting screen”.

<Step 4 (Distribution of comments for each website)>
The “result display screen” shows the distribution of the number of utterances of a group that is “positive” for the keyword (PC X), and the characteristics of the utterance contents in time series for each website belonging to the group (whether there are many positive statements) The graph 508 shown separately is displayed. On the “condition setting screen”, the Web site 509 having a lot of positive comments is selected. Here, the ellipse over a certain period in the graph 508 indicates that the larger the ellipse, the larger the number of statements over the period, and the thinner the ellipse color, the greater the number of positive comments. .

<Step 5 (words that occur on the site of interest)>
On the “result display screen”, the number of utterances for each word (for example, “design”, “function”, “price”, etc.) generated in the utterance on the Web site of interest selected in the previous step, and the utterance including the word A graph 510 is displayed for each of the utterance contents (a positive utterance, a negative utterance, or other moderate utterance). The bar in the graph 510 indicates that the lighter the color, the more positive the statement.

  The user performs a series of operations as described above, and finally uses keywords (for example, “design”, “function”, “price”, etc.) displayed on the “result display screen” in step 5 and keywords related to those words ( Potential needs for PC X) may be detected.

(Effect of embodiment)
As described above, according to the present embodiment, sentences are collected for a bulletin board and chat service on the Internet and a service that is a word-of-mouth communication among users, and the word of the word of review and the person related to the word of mouth are included in the word to be examined. Analyze it and provide it to users as a report. In addition, it is characterized in that a user can easily obtain a desired analysis result by setting in advance an analysis procedure for guiding the propagation of a Kramer, latent needs, reputation, and the like. This makes it possible to provide a service for notifying the actual state of word of mouth desired by companies and general consumers.

It is a system configuration figure showing a word-of-mouth information analysis system which realizes word-of-mouth analysis service of an embodiment of the present invention. It is a signal sequence diagram corresponding to the word-of-mouth information analysis system of Drawing 1 of an embodiment of the present invention. It is a figure which shows an example of the analysis procedure which the analysis report preparation module has set in the analysis report beforehand, when the user of the user computer of embodiment of this invention detects a Kramer. It is a figure which shows an example of the analysis procedure which the analysis report preparation module has set in the analysis report beforehand, when the user of the user computer of embodiment of this invention detects a Kramer. It is a figure which shows an example of the analysis procedure which the analysis report preparation module has set in the analysis report beforehand, when the user of the user computer of embodiment of this invention detects a latent need. It is a figure which shows an example of the analysis procedure which the analysis report preparation module has set in the analysis report beforehand, when the user of the user computer of embodiment of this invention detects a latent need. It is a system block diagram which showed an example of the word-of-mouth information analysis system which implement | achieves the conventional word-of-mouth analysis service. It is a signal sequence diagram corresponding to the word-of-mouth information analysis system of FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Speaker computer 102 Web server 103 Analysis server 104 Computer network 105 User computer 106 Browser 107 BBS software 108 Text collection module 109 Word extraction module 111 Analysis report creation module 112 Analysis report browsing software 201 Text analysis module 202 Speaker analysis module 203 Database 301, 303, 305, 307, 308, 309, 310, 311 Graph 501 Graph 503 Keyword 504, 506, 508, 510 Graph

Claims (12)

One or more first server systems connected to a computer network have a source that accepts information spoken by a speaker and publishes text sentences including the accepted information on the computer network A client system that collects and analyzes the text sentence from a second server system to generate a report, and that is connectable to the computer network is an information analysis system that displays information based on the report,
The first server system includes:
First analysis means for analyzing the collected text sentences at least in a body including the information, an identifier of a speaker of the body, an identifier of the source spoken by the speaker, and a speech period of the body;
Based on the analysis result obtained by the first analysis means, a second analysis is performed that counts at least the number of chronological utterances in the source for each utterer and specifies the characteristics of the utterance content. Means,
Report creation means for creating the report by graphing at least the analysis results obtained by the first analysis means and the second analysis means;
The client system is
Based on the report, displaying a graph showing the number of utterances for each speaker in chronological order, period selection means for receiving selection of a utterance period with a large number of utterances from the displayed graph;
Based on the report, a graph in which the speakers during the period selected by the period selection unit are grouped according to the number of the speakers and the characteristics of the contents of the speakers is displayed, and a specific group is displayed from the displayed graph. A group selection means for accepting the selection,
Based on the report, a graph showing a ratio of the total number of utterances of the speakers in the specific group selected by the group selection unit to the total number of utterances is displayed, and the specific group is displayed from the displayed graph. A rate selection means for accepting a selection for a rate of
Based on the report, the distribution of the number of utterances of the specific group selected by the ratio selection unit is displayed for each of the uttered sources in a chronological order according to the feature of the utterance content, and the displayed Source selection means for accepting selection of a specific source from the graph;
Information display means for displaying a graph indicating the number of utterances for each speaker in the source selected by the source selection means based on the report, according to the content of the utterance,
The report creation means creates the report in which the analysis results obtained by the first analysis means and the second analysis means are arranged in a graph format displayed by each means of the client system, An information analysis system characterized in that data settings for instructing the client system in order to sequentially execute each means of the client system are set in advance in the report. The information analysis system according to claim 1,
The feature of the content of the remark is a frequency indicating whether there are many positive remarks or negative remarks,
The specific group is a group having a large number of utterances by a speaker and having the most negative utterances as a feature of the utterance content.
The specific source is the source having the most negative comments as a feature of the content of the statements,
The information analysis system according to claim 1, wherein the content of the speech is at least a distinction between a positive speech and a negative speech.   3. A server system for an information analysis system, comprising: the first analysis unit, the second analysis unit, and the report creation unit included in the server system according to claim 1 or 2.   A client of an information analysis system comprising the period selection means, the group selection means, the ratio selection means, the source selection means, and the information display means included in the client system according to claim 1 or 2. system. One or more first server systems connected to a computer network have a source that accepts information spoken by a speaker and publishes text sentences including the accepted information on the computer network A client system that collects and analyzes the text sentence from a second server system to generate a report, and that is connectable to the computer network is an information analysis system that displays information based on the report,
The first server system includes:
First analysis means for analyzing the collected text sentences at least in a body including the information, an identifier of a speaker of the body, an identifier of the source spoken by the speaker, and a speech period of the body;
Second analysis means for extracting words from the body obtained by the first analysis means, and counting at least the number of chronological utterances in the source where the word is uttered for each extracted word When,
Report creation means for creating the report by graphing at least the analysis results obtained by the first analysis means and the second analysis means;
The client system is
Based on the report, displaying a graph showing the number of utterances of a predetermined word for each source in time series, period selection means for receiving selection of a utterance period with a large number of utterances from the displayed graph;
Based on the report, for the body containing the predetermined word during the period selected by the period selecting means, a graph in which the ratio of the number of utterances is grouped according to the utterance content for the predetermined word is displayed, and the display A group selection means for accepting a selection of a specific group from the generated graph;
Based on the report, the distribution of the number of utterances of the specific group selected by the group selection means is displayed in a time series for each of the uttered sources, and a graph showing the utterance content by characteristic is displayed. Source selection means for accepting selection of a specific source from the graph;
An information display means for displaying a graph showing the number of utterances for each word in the source selected by the source selection means based on the report, according to the content of the utterance of the body containing the word,
The report creation means creates the report in which the analysis results obtained by the first analysis means and the second analysis means are arranged in a graph format displayed by each means of the client system, An information analysis system characterized in that data settings for instructing the client system in order to sequentially execute each means of the client system are set in advance in the report. In the information analysis system of Claim 5,
The remark content is either positive or negative,
The feature of the content of the remark is a frequency indicating whether there are many positive remarks or negative remarks,
The specific source is the source having the most positive statements as a feature of the statement content,
The information analysis system according to claim 1, wherein the content of the speech is at least a distinction between a positive speech and a negative speech.   7. A server system for an information analysis system, comprising: the first analysis unit, the second analysis unit, and the report creation unit included in the server system according to claim 5 or 6.   7. A client system for an information analysis system, comprising: the period selection unit, the group selection unit, the source selection unit, and the information display unit included in the client system according to claim 5 or 6. One or more first server systems connected to a computer network have a source that accepts information spoken by a speaker and publishes text sentences including the accepted information on the computer network An information analysis method for an information analysis system in which a client system capable of collecting and analyzing the text sentence from a second server system to generate a report and connecting to the computer network displays information based on the report. ,
The first server system analyzes the collected text sentences at least in a body including the information, an identifier of a speaker of the body, an identifier of the source uttered by the speaker, and a speech period of the body. A first analysis step to:
Based on the analysis result obtained in the first analysis step, the first server system counts at least the number of chronological utterances in the source uttered for each utterer, and features of utterance contents A second analysis step for identifying
The first server system comprises a report creation step for creating the report by graphing at least the analysis results obtained in the first analysis step and the second analysis step;
The client system displays a graph showing the number of utterances for each speaker in chronological order based on the report, and a period selection step for receiving selection of a utterance period with a large number of utterances from the displayed graph;
Based on the report, the client system displays a graph in which the speakers during the period selected in the period selection step are grouped according to the number of utterances of the utterers and the characteristics of the utterance contents, and the displayed A group selection step for accepting selection of a specific group from the graph;
Based on the report, the client system displays a graph showing a ratio of the sum of the number of utterances of speakers in the specific group selected in the group selection step to the total number of utterances, and the displayed graph A ratio selection step for accepting a selection for the ratio of the specific group from:
Based on the report, the client system displays a graph showing the number of utterances of the specific group selected in the ratio selection step for each of the uttered sources in chronological order according to the feature of the utterance content. A source selection step for accepting a selection of the specific source from the displayed graph;
The client system includes an information display step of displaying a graph indicating the number of utterances for each speaker in the source selected in the source selection step based on the report, according to the content of the utterance,
In the report creation, the server system prepares the report in which the analysis results obtained in the first analysis step and the second analysis step are arranged in the form of a graph displayed in each step of the client system. A data analysis method for setting in the report in advance a data setting that instructs the client system to create and sequentially execute each step of the client system. The information analysis method according to claim 9,
The feature of the content of the remark is a frequency indicating whether there are many positive remarks or negative remarks,
The specific group is a group having a large number of utterances by a speaker and having the most negative utterances as a feature of the utterance content.
The specific source is the source having the most negative comments as a feature of the content of the statements,
The information analysis method according to claim 1, wherein the statement content is at least a distinction between a positive statement and a negative statement. One or more first server systems connected to a computer network have a source that accepts information spoken by a speaker and publishes text sentences including the accepted information on the computer network An information analysis method for an information analysis system in which a client system capable of collecting and analyzing the text sentence from a second server system to generate a report and connecting to the computer network displays information based on the report. ,
The first server system analyzes the collected text sentences at least in a body including the information, an identifier of a speaker of the body, an identifier of the source uttered by the speaker, and a speech period of the body. A first analysis step to:
The first server system extracts words from the body text obtained in the first analysis step, and at least for each extracted word, the number of chronological statements in the source where the word is spoken A second analysis step of counting
The first server system includes a report creation step of creating the report by graphing at least the analysis results obtained by the first analysis step and the second analysis step;
The client system displays a graph showing the number of utterances of a predetermined word for each source in time series based on the report, and receives a selection of a utterance period with a large number of utterances from the displayed graph. When,
Based on the report, the client system is a graph in which the ratio of the number of utterances is grouped according to the utterance content for the predetermined word with respect to the body including the predetermined word during the period selected in the period selection step. A group selection step for displaying and accepting a selection of a specific group from the displayed graph;
Based on the report, the client system displays a graph showing the number of utterances of the specific group selected in the group selection step according to the feature of the utterances in chronological order for each of the uttered sources. A source selection step for accepting a selection of the specific source from the displayed graph;
The client system includes an information display step for displaying a graph indicating the number of utterances for each word in the source selected in the source selection step based on the report, according to the content of the utterance,
In the report creation, the server system prepares the report in which the analysis results obtained in the first analysis step and the second analysis step are arranged in the form of a graph displayed in each step of the client system. A data analysis method for setting in the report in advance a data setting that instructs the client system to create and sequentially execute each step of the client system. In the information analysis method of Claim 11,
The remark content is either positive or negative,
The feature of the content of the remark is a frequency indicating whether there are many positive remarks or negative remarks,
The specific source is the source having the most positive statements as a feature of the statement content,
The information analysis method according to claim 1, wherein the statement content is at least a distinction between a positive statement and a negative statement.

Images