JP2005063241A - Information propagation measurement system in internet, and method therefor - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To realize measuring an effect of a topic having a great social interest in a variety of circumstances and locations, by measuring the spreading speed or intensity of the topic symbolized by predetermined words into the society. <P>SOLUTION: An analysis server 103 includes information source classification module 201, user identification module 202, date/time identification module 203 and application APL 205, in addition to the conventional modules. Also, in place of a propagation analysis module 110, a propagation measurement module 204 is included. As to Web servers in the Internet, information can be collected from a Web server 210 in an enterprise site and a Web server 211 which is an aggregation of personal pages, in addition to a Web server 101 of a BBS site. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention relates to an information propagation measurement system and method on the Internet, and in particular, by collecting various information published on the Internet and determining whether a topic is spreading widely in society. TECHNICAL FIELD The present invention relates to an information propagation measurement system and method on the Internet for performing information propagation analysis using a computer.

  Conventionally, the author of the article, the posting date and time, and the included word are analyzed for the target information such as the article on the Web server that can post information and publish on the Internet, and the word-of-mouth information that is the target information via the Internet As a technology to measure the situation of the propagation of word of mouth, there are the word-of-mouth propagation measurement method and word-of-mouth information analysis service (see Non-Patent Document 1) on the Internet, and the propagation is measured by comparing the values on two measurement dates. Yes.

  FIG. 1 is a diagram showing an example of a method for measuring word-of-mouth propagation on the Internet. In FIG. 1, a user computer 101 that refers to or posts word-of-mouth information, a Web server 102 that discloses word-of-mouth information, and the like and an analysis server 103 are connected via a computer network 104. The user computer 101 includes a browser 105 that can refer to or post word-of-mouth information, the Web server includes BBS software 106 that accepts posting and reference of word-of-mouth information, and the analysis server 103 includes a text collection module 107, A sentence analysis module 108, a word extraction module 109, and a propagation analysis module 110 are included.

  User computer 101 refers to or posts word-of-mouth information to BBS software 106 in Web server 102 using browser 105. The analysis server 103 collects text sentences including word-of-mouth information from the BBS software 106 in the Web server 102 using the sentence collection module 107. In the sentence analysis module 108, the collected text sentences are divided into a text including word-of-mouth information, information for identifying a contributor who is a creator (poster name), and date information indicating a posting date and time.

  The text including the word-of-mouth information is transmitted to the word extraction module 109, divided into words by morphological analysis or the like, and returned to the sentence analysis module 108. The sentence analysis module 108 transmits the poster name, the posting date and the extracted word (noun) to the propagation analysis module 110 for each posted sentence. The propagation analysis module 110 measures the word-of-mouth effect for each word based on the time-series data of the number of sentences including the word and the number of posted persons.

  FIG. 2 shows an example of a signal sequence diagram for FIG.

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 conventional method for measuring word-of-mouth propagation, although it is possible to count bulletin boards and chats on the Internet, information on other Web servers is not taken into consideration, and the number of people accessed within a certain period can be predicted or estimated. There is a problem in that it is inaccurate to measure the propagation of information widely throughout society because it is impossible to accurately count the number of people who access within a certain period.

  In the present invention, for information including a word that symbolizes a certain topic, the information source is classified as a counting means for a wide range of Web servers, and an accurate number of people who access within a certain period for each information source type is targeted. Introducing a counting means, a means for measuring the date of transmission of information, a means for counting information viewers, and a means for counting the propagation of information, and measuring the propagation of information on the Internet .

  Specifically, the information propagation measurement system on the Internet according to claim 1 is one or more Internet sites for publishing target information having one or more sentences, and access to the disclosed target information. One or more client computers including one or more public servers including history recording means for recording history and means for writing target information to the public servers via the one or more Internet sites connected to the Internet And collecting means for collecting target information from a public server via the one or more Internet sites connected to the Internet, extracting date information and creator of the target information from the collected target information, Analyzing and extracting means for analyzing a sentence and extracting words from the sentence, and one or more predetermined words An analysis server including a totaling unit that counts the number of target information and the number of creators who created target information including a predetermined word based on date information for each predetermined word and calculates a penetration degree It is characterized by that.

  The invention according to claim 2 is the information propagation measurement system on the Internet according to claim 1, wherein the penetration degree is obtained by dividing the square of the number of creators totaled by the number of target information aggregated. It is set as the value obtained.

  According to a third aspect of the present invention, in the information propagation measurement system on the Internet according to the second aspect, the counting means totals the number of pieces of target information including one or more predetermined words for each public server. , And multiplying by a predetermined weight value assigned in advance to each of the public servers.

  According to a fourth aspect of the present invention, in the information propagation measurement system on the Internet according to the first aspect, the analysis server receives an access history from a public server via the Internet, and receives one or more predetermined words. It further includes a reference counting means for counting the number of referees who have accessed the target information, based on the access date and time, and the penetrance is a value obtained by adding the number of referrals to the square of the total number of creators. A value obtained by dividing by the total number of pieces of target information is characterized.

  According to a fifth aspect of the present invention, in the information propagation measurement system on the Internet according to the second, third, or fourth aspect, the tabulating means sets the number of target information including one or more predetermined words to one. Alternatively, the calculation is performed for each of a plurality of predetermined words, and is calculated by multiplying each of the one or more predetermined words by a predetermined weight value assigned in advance.

  According to a sixth aspect of the present invention, in the information propagation measurement system on the Internet according to the first aspect, the analysis server uses an asymptotic function that gradually increases to an infinite value when the input value is increased to an infinite value. Is further input as an asymptotic function input value to obtain an output value asymptotic to a certain value.

  According to a seventh aspect of the present invention, in the information propagation measurement system on the Internet according to the first aspect, the penetrance is a square of a value obtained by accumulating the number of creators counted from a predetermined date and time, from the predetermined date and time. It is characterized in that it is a value obtained by dividing the total number of target information by the accumulated value.

  According to an eighth aspect of the present invention, in the information propagation measurement system on the Internet according to any one of the first to seventh aspects, the analysis server further includes means for calculating a change amount of the penetration degree with respect to a predetermined time interval. It is characterized by that.

  The invention according to claim 9 is that the analysis server further includes means for ordering predetermined words in descending order of penetration in the information propagation measurement system on the Internet according to any one of claims 1 to 7. Features.

  The invention according to claim 10 is the information propagation measurement system on the Internet according to any one of claims 1 to 7, and the occurrence date and time of a plurality of events for a product or event symbolized by a predetermined word, It further comprises event effect calculation means for measuring the propagation degree based on the word penetration degree and calculating the effect for each of the plurality of events based on the propagation degree.

  The invention according to claim 11 is an information propagation measurement method on the Internet, in which one or a plurality of public servers publish target information having one or a plurality of sentences on one or a plurality of Internet sites. A step of recording an access history to the disclosed target information on the history recording means, and one or a plurality of client computers connect to the Internet and write the target information to a public server via one or a plurality of Internet sites. Step, the analysis server connects the collection means to the Internet, collects the target information from the public server via one or more Internet sites, and creates the date information and the target information from the target information collected by the analysis extraction means Extract the person, analyze the sentence of the target information and extract the word from the sentence, In addition, the counting means counts the number of pieces of target information including one or more predetermined words and the number of creators who created the target information including predetermined words based on date information for each predetermined word. And a step of calculating.

  In the present invention, by measuring the propagation of information on the Internet for information including a word that symbolizes a topic, the speed and strength of the spread of the topic symbolized by a word to society is measured. In addition, it is possible to extract topics that are highly interested in society, rank topics, measure the effects of advertisements, campaigns, and press releases.

  As described above, the present invention includes one or more Internet sites for publishing target information having one or more sentences, and history recording means for recording an access history to the disclosed target information. One or more public servers, one or more client computers including means for connecting to the Internet and writing the target information to the public server via one or more Internet sites, and one or more connected to the Internet Collecting means for collecting target information from public servers via the Internet site, and extracting date information and the creator of the target information from the collected target information, analyzing the sentence of the target information and extracting words from the sentence Analyzing and extracting means, the number of pieces of target information including one or more predetermined words, and target information including the predetermined words Introducing an analysis server that includes an aggregation means for calculating the degree of penetration by counting the number of creators who have created for each predetermined word based on date information, and measuring the propagation of information on the Internet By measuring the speed or strength of the topic that is symbolized by a certain word to the society, the effect on the topic that is highly interested in society can be measured in various situations.

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

  FIG. 3 is a diagram showing an example of the information propagation measurement system of the present invention. In FIG. 3, the information server classification module 201, the user identification module 202, the date and time identification module 203, and the usage APL 205 are added to the analysis server 103 in addition to the conventional one, and a propagation measurement module 204 is included instead of the propagation analysis module 110. . As a Web server on the Internet, in addition to the Web server 101 of the BBS site, information can also be collected from the Web server 210 of the corporate site and the Web server 211 that is a collection of personal pages.

  In collecting information from the BBS site, the analysis server 103 uses the text collection module 107 to collect text sentences published on the Web server 101 that is the BBS site, and notifies that it has been collected from the Web server of the BBS site. The collected text sentences are transmitted to the information source classification module 201. The information source classification module 201 divides the text of the information and the information (user name) for identifying the poster and the posting date and time from the text sentence according to the preset information processing procedure of the BBS site. The word extraction module 109 is transmitted, divided into words by morphological analysis or the like, and returned to the sentence analysis module 108. The information source classification module 201 transmits the user name and date information to the propagation measurement module 204 for each extracted word (noun).

  In collecting information from a corporate site, the analysis server 103 uses the text collection module 107 to collect text sentences published on the Web server 210 that is a corporate site. The information source classification module 201 notifies the fact that it has been collected from a Web server that is a corporate site and transmits the collected text text to the information source classification module 201. The site information is identified as a single user, and date information is extracted from the HTTP header portion of the text information. The text of the information is transmitted to the word extraction module 109, divided into words by morphological analysis or the like, and returned to the sentence analysis module 108. The information source classification module 201 transmits the name of the user who is the owner of the information, the creation date of the information, and the extracted word (noun) to the propagation measurement module 204.

  In collecting information from personal pages, the analysis server 103 uses the text collection module 107 to collect text sentences published on the Web server 211, which is a collection of personal pages, and collects the personal pages. The collected text sentences are notified to the collection from a certain Web server, and the collected text sentences are transmitted to the information source classification module 201. The information source classification module 201 identifies the owner of the information under the URL for identifying the individual as a single user in accordance with the processing procedure of the personal page information set in advance, and the date information from the header portion of the text information in HTTP. The text of the information is transmitted to the word extraction module 109 and returned to the sentence analysis module 108 that divides the information into words by morphological analysis or the like.

  The information source classification module 201 transmits the name of the user who is the owner of the information, the creation date of the information, and the extracted word (noun) to the propagation measurement module 204.

  The propagation measurement module 204 calculates the penetration degree using the words collected from the information source classification module 201 and the user name and date given to the words, and outputs the calculation result according to the request from the usage APL 205 and the usage.

  The usage APL 205 summarizes and reports the aggregation results according to various applications required by users such as companies based on various calculation results and information including the penetration degree obtained from the propagation measurement module 204, Or make future predictions. In the present embodiment, the use APL 205 is described as being included in the analysis server 103, but in other embodiments, it may be included in a client computer such as a company or another server.

  FIG. 4 shows an example of a signal sequence diagram for FIG. As shown in FIG. 4, the text collection module 107 and the information source classification module 201 access all the Web servers 101, 201, and 211, collect information, and classify the information sources. Similarly, the sentence analysis module 108 and the word extraction module 109 analyze the collected information and extract words. However, what is identified by the user identification module 203 normally handles only information from the corporate web server 210 and the personal web server 211.

  FIG. 5 is an example of a BBS screen on the Web. On the BBS site, messages such as those shown in Fig. 5 are usually posted, and these are the targets of collection. On the Web screen, the posted text is displayed with the name of the contributor and the date and time of posting. ing.

  FIG. 6 is an example of information on a corporate Web server. As shown in FIG. 6, information output in normal HTTP includes an HTTP header and an HTML text. In this embodiment, information is collected from the HTML text, and date information is acquired from the HTTP header. User identification information is extracted from the URL portion of the company site.

  FIG. 7 is an example of information on the personal Web server. Basically, various information is collected or extracted in the personal server as well as in the enterprise server, but the user identification information is extracted mainly using the lower part of the URL of the site.

  FIG. 8 is an example of word extraction using morphological analysis in the word extraction module 109. As shown in FIG. 8, the input sentence is subjected to morphological analysis, word extraction is performed, and finally an extraction result is obtained.

In the following, some examples relating to the collection of information will be described.
(First embodiment)
The penetrance calculation method for the topic X in a certain period according to the first embodiment of the present invention is as follows.

(1) A keyword characterizing the topic X is a word Y and a word Z.
(Example: Topic X = “Tama-chan”, Word Y = “Tama-chan”, Word Z = “Tama-chan”)
(2) Count the number of articles including the word Y + the number of articles including the word Z, and count the number of authors of the accurate articles of the article including the word Y and the article including the word Z.
(3) An example of counting results is shown in the following table.

  Number of articles on the above topic in a week:

  Number of authors for the above topic for a week (the top is the exact number of authors for the day, the bottom is the exact number of authors throughout the cumulative period):

  Here, an example is shown in which site A is the same person every day and site B is a different person every day.

  Number of viewers for this article counted from the logs recorded on the Web servers of Site A and Site B (the upper row is the current day, the lower row is cumulative):

  Here, as in the above table, it is assumed that the site A is the same person every day and the site B is a different person every day, and accurate access users may be counted in combination with Cookie or the like.

  Constant values for sites and topics:

  Here, the constant value of the site can be calculated using the number of registered users and the number of accesses of the site, and the constant value for the topic can also be calculated from the ranking of search keywords such as a search engine.

(Second embodiment)
In the second embodiment of the present invention, the penetrance is calculated by the method shown in claim 2. Specifically, it is as follows. (Example of penetrance calculation for 7 days measurement period)
(2-1) Total number of articles on site A for 7 days + total number of articles on site B for 7 days = 140
(2-2) Total number of authors for site A for 7 days (= 5) + total number of authors for site B for 7 days (= 35) = 40
(2-3) Penetrance = square of (2-1) / (2-2) = 1600/140 = 11.4
It becomes.

  In the above description, the penetrance is calculated for the entire seven days. However, it is also possible to see how the penetrance changes every day by applying the calculation method of this embodiment every day. The results of calculating the penetrance from the first day to the seventh day under the above conditions are as follows.

(Third embodiment)
In the third embodiment of the present invention, the penetrance is calculated by the method shown in claim 3. Specifically, it is as follows. (Example of penetrance calculation for 7 days measurement period)
(3-1) Number of articles for seven days at site A (= 70) × constant value (= 1/10) + number of articles for seven days at site B (= 70) × constant value (= 1) = 77
(3-2) Permeability = (2-1) squared / (3-1) = 1600/77 = 20.8
It becomes.

(Fourth embodiment)
In the fourth embodiment of the present invention, the degree of penetration is calculated by the method shown in claim 4. Specifically, it is as follows. (Example of penetrance calculation for 7 days measurement period)
(4-1) The square of the total number of authors of site A and site B (= 1600) + the number of referrals of site A (= 160) + the number of referrals of site B (= 2450) = 4210
(4-2) Penetrance = (4-1) / (2-1) = 4210/140 = 30.1
It becomes.

(5th Example)
In the fifth embodiment of the present invention, the permeability is calculated by the method shown in claim 5. Specifically, it is as follows.
(5-1) Penetrance = (2-3) × 10 = 11.4 × 10 = 114
It becomes.

(Sixth embodiment)
In the fifth embodiment of the present invention, the permeability is calculated by the method shown in claim 5. Specifically, it is as follows. The following equation is asymptotic to 1.
(6-1) Corrected penetrance = tanh (penetration / 100)

(Seventh embodiment)
In the seventh embodiment of the present invention, the permeability is calculated by the method shown in claim 7. Specifically, it is as follows. This example is similar to the daily penetration of Example 2, but is characterized in that the accumulated penetration is calculated. Therefore, the penetrance on the seventh day is the same as the penetrability calculated for the entire seven days. The calculation example of the penetration degree of the present embodiment is as follows.

(Eighth embodiment)
In the eighth embodiment of the present invention, the permeability is calculated by the method shown in claim 8. Specifically, it is as follows.

  As described above, since the difference from the 6th to the 7th is 1 for the topic I, 4 for the topic J, and 1 for the topic K, the topic J is extracted as a topic of strong interest on the 7th.

(Ninth embodiment)
In the ninth embodiment of the present invention, the permeability is calculated by the method shown in claim 9. Specifically, it is as follows.

  As described above, the ranking of the penetration degree on the 7th is Topic K> Topic I> Topic J.

(Tenth embodiment)
Similarly, in the tenth embodiment of the present invention, such an ordering is performed based on the occurrence of an event such as a press release date or an advertisement date, so that the penetration of a word symbolizing a product or an event is written together. The degree can be measured and the effect of events such as press releases or advertisements can be calculated. Thus, the result of the Example which performed the penetration degree ordering for every event is as follows.

  As described above, by measuring the degree of propagation such as the speed of spread of goods or events symbolized by a word to society, the timing of spread by referring to the press release date or the date of advertisement, etc., by event A The effect was 5 for the day, +5 for the next day, the effect for event B was 5 for the day, -2 for the next day, and the effect for event C was 20 for the day and +1 for the next day. Therefore, it is determined that event A and event C are effective, but event B is not effective.

  Therefore, in the present invention, for information including a word symbolizing a certain topic, a counting means for a wide range of Web servers, a counting means for a unique number of people within a certain period, and an information viewer Introducing a target counting means and a means to count the propagation of information, and measuring the propagation of information on the Internet, the speed of spreading to the society of the topic symbolized by a word, the strength of spreading By measuring the degree of interest, it is possible to extract topics that are highly interested in society, rank topics, measure the effects of advertisements, campaigns, and press releases.

It is the figure which showed an example of the conventional word-of-mouth propagation measurement method on the internet. It is a figure which shows an example of the signal sequence diagram with respect to FIG. It is a figure which shows an example of the information propagation measurement system of this invention. It is a figure which shows an example of the signal sequence diagram with respect to FIG. It is a figure which shows an example of the BBS screen on Web. It is a figure which shows an example of the information on a corporate web server. It is a figure which shows an example of the information on a personal type | system | group Web server. It is a figure which shows an example of the word extraction using the morphological analysis in the word extraction module 109. FIG.

Explanation of symbols

101 User computer 102, 210, 211 Web server 103 Analysis server 104 Computer network 105 Browser software 106 BBS software 107 Text collection module 108 Text analysis module 109 Word extraction module 110 Propagation analysis module 201 Information source classification module 202 User identification module 203 Identification module 204 Propagation measurement module 205 APL module used

Claims (11)

One or more public servers including one or more Internet sites for publishing target information having one or more sentences, and history recording means for recording an access history to the disclosed target information;
One or more client computers including means for connecting to the Internet and writing the target information to the public server via the one or more Internet sites;
A collection unit that connects to the Internet and collects the target information from the public server via the one or more Internet sites, and extracts date information and a creator of the target information from the collected target information; Analyzing and extracting means for analyzing a sentence included in the target information and extracting words from the sentence, the number of the target information including one or more predetermined words, and creating the target information including the predetermined words An information propagation measurement system on the Internet, comprising: an analysis server including a totaling unit that counts the number of creators for each predetermined word based on the date information and calculates a penetration degree.   2. The information propagation on the Internet according to claim 1, wherein the penetration degree is a value obtained by dividing the square of the total number of creators divided by the total number of target information. Measuring system.   The counting means counts the number of the target information including the one or more predetermined words for each of the public servers, and counts by multiplying by a predetermined weight value assigned in advance to each of the public servers. The information propagation measurement system on the Internet according to claim 2. The analysis server receives the access history from the public server via the Internet, and counts the number of reference users who have accessed the target information including the one or more predetermined words based on the access date and time. It further includes a referrer counting means,
The penetration degree is a value obtained by dividing a value obtained by adding the number of reference users to the square of the number of the collected creators by the number of the collected target information. Item 3. The information propagation measurement system on the Internet according to Item 1.   The counting means counts the number of the target information including the one or more predetermined words for each of the one or more predetermined words, and preliminarily adds each of the one or more predetermined words. 5. The information propagation measurement system on the Internet according to claim 2, 3 or 4, wherein the total is multiplied by a predetermined weight value assigned.   The analysis server uses an asymptotic function that increases an input value to be infinite and uses an asymptotic function that gradually approaches a constant value, and inputs the penetration degree as an input value of the asymptotic function to obtain an output value that asymptotically approaches a certain value. The information propagation measurement system on the Internet according to claim 1, further comprising:   The penetration degree is a value obtained by dividing the square of a value obtained by accumulating the number of creators aggregated from a predetermined date and time by a value obtained by accumulating the number of target information aggregated from the predetermined date and time. The information propagation measurement system on the Internet according to claim 1.   8. The information propagation measurement system on the Internet according to claim 1, wherein the analysis server further includes means for calculating a change amount of the penetration degree with respect to a predetermined time interval.   8. The information propagation measurement system on the Internet according to claim 1, wherein the analysis server further includes means for ordering the predetermined words in descending order of penetration.   Measures the degree of propagation based on the occurrence date and time of a plurality of events for a product or event symbolized by a predetermined word and the penetration degree of the word, and calculates the effect for each of the plurality of events based on the degree of propagation The information propagation measurement system on the Internet according to claim 1, further comprising event effect calculation means for performing the operation. A step of causing one or a plurality of public servers to publish target information having one or more sentences on one or a plurality of Internet sites, and recording history of access to the disclosed target information in a history recording unit; ,
One or more client computers connecting to the Internet and writing the object information to the public server via the one or more Internet sites;
An analysis server connects the collection means to the Internet, collects the target information from the public server via the one or more Internet sites, and causes the analysis extraction means to collect date information and the target from the collected target information The creator of the information is extracted, the sentence included in the target information is analyzed, words are extracted from the sentence, and the number of the target information including one or more predetermined words is counted in the counting unit, and the predetermined number And the number of creators who created the target information including the word of the number of the predetermined information for each predetermined word based on the date information, and calculating a penetration degree. Propagation measurement method.

Images