JP2004139376A - Word-of-mouth propagation measuring method on the internet - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To overcome the problem that the effect as word-of-mouth communication can not be determined because it is not analyzed about how much word-of-mouth communication is recognized widely in the conventional word-of-mouth analysis. <P>SOLUTION: By this method, users can contribute texts on the Internet, and the texts are collected from the object of a bulletin board system (BBS) and chat which can specify the persons and the time and date concerning contributed texts and references and the resembled services which are the media of word-of-mouth communication between users. A mechanism analyzes the sentences by the number of persons and the time series using the words to investigate (word-of-mouth) is introduced, and whether the words (word-of-mouth) are going to expand or to decrease on the Internet is determined by this method. <P>COPYRIGHT: (C)2004,JPO

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a method for recognizing word-of-mouth communication on the Internet, and more particularly, to collect word-of-mouth information from general users exchanged on the Internet, and to determine whether word-of-mouth of a certain topic is spreading or converging. By making the determination, word-of-mouth analysis using the Internet is performed.
[0002]
[Prior art]
2. Description of the Related Art Conventionally, as a mechanism for collecting and analyzing word-of-mouth information on the Internet, there is an analysis technology that applies a web clipping service and text mining, for example, a true teller of Nomura Research Institute (Non-Patent Document 1).
[0003]
Here, text information is widely collected from the Internet, words are cut out by morphological analysis, and words extracted by multivariate analysis or statistical methods are analyzed for quantitative indices and their time series. (In TrueTeller V.2, the analysis functions include word frequency ranking / drilldown analysis, analysis of main topics, characteristic word analysis, mapping, detailed text display, group attribute analysis, time series monitor, scoring, report Is provided.)
[0004]
FIG. 1 shows an example of a conventional word-of-mouth analysis, and FIG. 3 shows an example of a signal sequence diagram for FIG. 1 and 3, a user computer 101 that refers to or posts word-of-mouth information, a Web server 102 on which word-of-mouth information and the like are published, and an analysis server 103 are connected to a computer network 104. The user computer 101 has a browser 105 capable of referring to or posting word-of-mouth information, the Web server has BBS software 106 for posting and receiving word-of-mouth information, the analysis server 103 has a sentence collection module 107, and a word extraction module. A module 108 and a word analysis module 109 are provided.
[0005]
The user computer 101 refers to or posts word-of-mouth information to the BBS software 106 in the Web server 102 using the browser 105. The analysis server 103 uses the text collection module 107 to collect text text including word-of-mouth information from the BBS software 106 in the web server 102. The collected text sentences are divided into word units such as business cards by a method such as morphological analysis in the word extraction module 108. The extracted words are totaled for each word type in the word analysis module 109, and the correlation value between the words is counted by the appearance frequency or a statistical method.
[0006]
[Non-patent document 1]
Nomura Research Institute, "What is true teller?"
[0007]
[Problems to be solved by the invention]
Traditional word-of-mouth analysis shows how words appearing on the Internet have changed, and does not take into account cases where specific people intentionally wrote or read articles. There has been a problem that how widely the word has been recognized as word of mouth has not been analyzed, and the effect as word of mouth cannot be determined.
[0008]
[Means for Solving the Problems]
According to the present invention, texts can be collected on a bulletin board (BBS) that allows a user to post a text on the Internet and can specify a person and a date and time with respect to a posted text and a reference, and a chat and a similar service such as a word-of-mouth communication between users. It also introduces a mechanism to analyze the number of people who posted using the word (word of mouth) to be examined and its time series, and determines whether the word (word of mouth) is spreading or declining on the Internet. And
According to the present invention, sentences are collected for bulletin boards and chats on the Internet and similar services serving as word-of-mouth media between users, and it is determined whether the word (word of mouth) to be examined is spreading or declining on the Internet. By doing so, it is possible to determine the word-of-mouth effect of the word.
[0009]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, the present invention will be described with reference to the drawings.
FIG. 2 shows an example of word-of-mouth analysis according to the present invention, and FIG. 4 shows an example of a signal sequence diagram for FIG. 2 and 4, a sentence analysis module 201 and a propagation analysis module 202 are added to the analysis server 103.
[0010]
The user computer 101 refers to or posts word-of-mouth information to the BBS software 106 in the Web server 102 using the browser 105. The analysis server 103 uses the text collection module 107 to collect text text including word-of-mouth information from the BBS software 106 in the web server 102. In the sentence analysis module 201, the collected text sentence is divided for each sentence of the sentence, and is divided into a text including word-of-mouth information, information (poster name) for identifying the contributor, and a posting date and time. The text including the word-of-mouth information is transmitted to the word extraction module 108, divided into words by morphological analysis or the like, and returned to the text analysis module 201. The sentence analysis module 201 transmits, to the propagation analysis module 202, the name of the poster, the date and time of the posting, and the extracted words (nouns) for each of the posted sentences. The propagation analysis module 202 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.
[0011]
FIG. 5 is an example of a BBS screen on the Web. On the screen on the Web, the posted text is displayed with the poster name and the posting date and time added.
[0012]
FIG. 6 is an example of word extraction using morphological analysis in the word extraction module 108. In this example, the input sentence is "Which one is better, Kana Preparatory School or Otsu Preparatory School?" This input sentence is analyzed by morphological analysis into "Noun-proper noun-personal name-surname", "preparatory school noun-general", "to particle-parallel particle", as shown in FIG. Next, as shown in (C), word extraction is performed based on a correction rule of “if nouns are continuous, one noun (compound)” and an extraction rule of “extract only nouns other than pronouns”. As a result, as shown in (D), two words, “Ko Preparatory School” and “Otsu Preparatory School”, were extracted.
[0013]
FIG. 7 shows an example of processing in the text analysis module 201, and FIG. 7B shows a processing result of the input text shown in FIG. For example, as shown in the first line of (B), the processing result of the article number 1 of the BBS number 1001 in (A) is the sentence number 1001-1, the author user1, the posting date 2002-03-22, and the word (Nouns) Kou Preparatory School and Otsu Preparatory School. The same applies to the article numbers 2 to 6.
[0014]
FIG. 8 is an example of a process in the propagation analysis module 202. The processing result of the sentence analysis module 201 shown in FIG. 7B is converted into time-series data for each word as shown in FIG. For example, a sentence containing the word “Kou Preliminary School” was posted by one contributor of user1 on March 22, 2002, the sentence number is 1001-1, and the number of posted articles is one. Posted on March 23, 2002 by two contributors, user3 and user2, their text numbers are 1001-2, 1001-3, and 1001-6, and the number of posted articles is three. The same applies to Otsu Preparatory School. (B) shows an example in which the propagation is evaluated based on the data of (A). Here, three types of evaluations (1) to (3) are performed.
[0015]
(1) Evaluation based on the sum of the number of contributors and the number of posted articles Regarding the Kobe Preparatory School, the sum of the number of contributors and the number of contributed articles was 2 on March 22, 2003, with 1 person + 1 article. On March 23, two people + three = five, so it is judged that Kou Preliminary School has spread. Regarding Otsu Preparatory School, one person + 1 = 2 on March 22, 2003, but 3 + 3 = 6 on March 23, 2003. Judge. Judging from the comparison of the figures, it is judged that the Otsu Preparatory School has spread more widely than the Ko Preparatory School.
[0016]
(2) Evaluation based on the ratio of the number of posts to the number of articles The ratio of the number of posts to the number of articles to be posted in the preliminary school was 1/1/1 on March 22, 2003. On March 23, 2003, 2/3 people = 0.6, so it is determined that propagation has attenuated. Regarding the Otsu Preparatory School, one person / one = 1 on March 22, 2003, and three people / three = 1 on March 23, 2003. Comparing the transmissions between the Ko-ko Preparatory School and the Ko-O Preparatory School, it is judged that the Ko-O Preparatory School has spread more widely than the Ko Ko Preparatory School.
[0017]
(3) Evaluation Party prep by with the articles the ratio of the number of the square of the (post number) is, the square with the articles the ratio of the number of (post number) is, on March 22, 2003 (1 person) ² / what it had been a one = 1, to March 23, 2003 from the fact that became a 2/3 = 1.3 (2 ^people), it is determined that the propagation. For Otsu prep school, on March 22, 2003 what it was ^(one) 2/1 = 1, to March 23, 2003 became (3 ^people) 2/3 = 3 Therefore, it is determined that the propagation has occurred. Judging from the comparison of the figures, it is judged that the Otsu Preparatory School has spread more widely than the Ko Preparatory School.
[0018]
FIG. 9 shows another example of the processing in the propagation analysis module 202. FIG. 9A is time-series data for each word, and indicates the posting date, the number of speakers, and the number of appearances for words A to F. Here, as shown in FIG. 9B, four types of evaluations (1) to (4) are performed.
[0019]
(1) Evaluation by "sum of the number of speakers and the number of appearances" in the change from measurement day 1 to measurement day 2 (claim 2)
For example, for word A, the sum of the number of speakers and the number of appearances has decreased from 4 + 4 = 8 on measurement day 1 to 2 + 2 = 4 on measurement day 2, so that the number of propagation = -4. The number of propagation is the difference between the value on measurement day 2 and the value on measurement day 1. Since the number of propagation is -4, it is determined that the propagation is attenuated. Since the propagation number of word B is 4, it is determined that the word B has propagated. The determination is similarly performed for the words C to F.
[0020]
(2) Evaluation by "ratio of the number of speakers to the number of appearances" in the change from measurement day 1 to measurement day 2 (claim 3)
For example, for word A, the ratio of the number of speakers to the number of appearances is 4/4 = 1 on measurement day 1 and 2/2 = 1 on measurement day 2, so that the number of propagation = 0. This determines that the propagation is the current status. The same determination is made for words BF.
[0021]
(3) Evaluation based on "ratio of the square of the number of speakers to the number of appearances" in the change from measurement day 1 to measurement day 2 (claim 4)
For example, for the word A, the ratio between the number of speakers and the number of appearances decreases from 16/4 = 4 on the measurement day 1 to 4/2 = 2 on the measurement day 2, and the number of propagation is −2. Is judged to have decreased. The same determination is made for words BF.
[0022]
(4) Sum of the above propagation numbers (Claim 5)
For example, for the word A, since the sum of the propagation numbers in the above (1) to (3) is (−4) +0 + (− 2) = − 5, it is determined that the propagation is attenuated. Since the word B is 4 + 0 + 2 = 6, it is determined that the word B has been propagated. The determination is similarly performed for the words C to F.
[0023]
8 and 9, one of “the number of speakers”, “the number of appearances”, “the ratio of the number of speakers to the number of appearances”, and “the ratio of the square of the number of speakers to the number of appearances” The value calculated by multiplying the above by a constant value may be used to determine whether word-of-mouth propagation is spreading or attenuating, maintain the current state, or determine the strength of word-of-mouth propagation (claim Item 6).
[0024]
As described above, the invention made by the inventor has been specifically described based on the embodiment. However, the present invention is not limited to the embodiment, and can be variously modified without departing from the gist of the invention. Needless to say,
[0025]
【The invention's effect】
As described above, according to the present invention, texts are collected for bulletin boards and chats on the Internet and similar services that serve as word-of-mouth media between users, and the number of people who posted using words (word-of-mouth) to be examined and their time series Introduce a mechanism to analyze the word, and determine whether the word (word of mouth) is spreading or declining on the Internet, and it is possible to determine the effect of the word as word of mouth .
[Brief description of the drawings]
FIG. 1 shows an example of a conventional word-of-mouth analysis.
FIG. 2 shows an example of word-of-mouth analysis of the present invention.
FIG. 3 is an example of a signal sequence diagram of a conventional word-of-mouth analysis.
FIG. 4 is an example of a signal sequence diagram for word-of-mouth analysis of the present invention.
FIG. 5 is an example of a BBS screen on the Web.
6 is an example of word extraction using morphological analysis in the word extraction module 108. FIG.
FIG. 7 is an example of processing in a sentence analysis module 201;
FIG. 8 is an example of processing in a propagation analysis module 202;
FIG. 9 is another example of the processing in the propagation analysis module 202;
[Explanation of symbols]
101: User computer, 102: Web server, 103: Information distribution server, 104: Computer network, 105: Browser, 106: BBS software, 107: Sentence collection module, 108: Word extraction module, 109: Word analysis module, 201 … Sentence analysis module, 202… propagation collection module

Claims (6)

A Web server that is connected to the Internet and has a mechanism for accepting posting of texts and adding a character string for identifying the date and time and the speaker to publish the texts, and a mechanism for collecting texts on Web pages connected to the Internet. In a network composed of analysis servers equipped with
The analysis server extracts words included in the sentence, a date and time when the sentence was posted, and a character string identifying a speaker of the sentence from the collected sentence,
The number of sentences including a word is defined as the number of occurrences of the word, and the number of occurrences and the number of speakers for each word are counted for each word,
Word-of-mouth communication that determines whether word-of-mouth communication is spreading or declining, keeps the status quo, or determines the strength of word-of-mouth communication, depending on the change in the number of appearances and the number of speakers of the word to be examined for each date. Permeability measurement method. In the word-of-mouth penetration measuring method according to claim 1,
Measure the “sum of the number of speakers and the number of appearances” for each date of the word you want to check, and if the “sum of the number of speakers and the number of appearances” increases during the measurement day, the spread of word-of-mouth is expanding and decreasing. If so, the word-of-mouth propagation is attenuating, and if it has not changed, word-of-mouth propagation is judged to be the status quo, or word-of-mouth propagation is determined by the difference between the "sum of the number of speakers and the number of appearances" during the measurement day. Word-of-mouth penetration measurement method characterized by measuring the size of a word-of-mouth. In the word-of-mouth penetration measuring method according to claim 1,
Measure the "ratio of the number of speakers to the number of appearances" for each date of the word you want to check, and if the "ratio of the number of speakers to the number of appearances" increases during the measurement day, the spread of word-of-mouth is spreading and decreasing. If so, word-of-mouth propagation is decreasing, and if it has not changed, word-of-mouth propagation is determined to be the status quo, or word-of-mouth propagation is determined by the difference between the ratio of the number of speakers and the number of appearances during the measurement day. Word-of-mouth penetration measurement method characterized by measuring the size of a word-of-mouth. In the word-of-mouth penetration measuring method according to claim 1,
Measure the "ratio of the number of speakers to the number of appearances" for each date of the word you want to check, and if the "ratio of the number of speakers to the number of appearances" increases during the measurement days, word of mouth spreads Yes, if it is decreasing, word-of-mouth propagation is decreasing, and if it has not changed, it is judged that word-of-mouth propagation is the status quo, or the ratio of the square of the number of speakers to the number of appearances during the measurement day The word-of-mouth penetration measurement method characterized in that the magnitude of word-of-mouth propagation is measured based on the difference of the word-of-mouth communication. In the word-of-mouth penetration measuring method according to any one of claims 2 to 4,
Word-of-mouth by combining each change of "sum of the number of speakers and the number of appearances", "ratio of the number of speakers and the number of appearances", and "ratio of the square of the number of speakers and the number of appearances" A word-of-mouth penetration measurement method characterized by determining whether propagation of a word is spreading or attenuating, maintaining the status quo, or determining the strength of word-of-mouth propagation. The word-of-mouth penetrance measurement method according to any one of claims 2 to 5,
Constant value for one or more of "number of speakers", "number of appearances", "ratio of number of speakers and number of appearances", "ratio of number of speakers and number of appearances" A word-of-mouth penetration measurement method characterized by determining whether word-of-mouth propagation is spreading or attenuating, maintaining the current state, or determining the strength of word-of-mouth propagation using a value calculated by multiplying word-of-mouth communication.

Images