CN106471494B - Program, apparatus and method for analyzing effect of promotion site on user's psychological state transition - Google Patents

Abstract

There is provided a program for analyzing the effect of a promotion site for a product or service on a target keyword, which effect acts on a user's mental state transition, based on a set of comment texts transmitted by each contributor. The program causes the computer to function as: a psychological state determiner that outputs a psychological state of psychological keywords included in the comment text transmitted by each contributor; a score memory that stores, for each promotion site address, a score set for each psychological state; and a score updater that instructs the score memory to increase the score of the mental state of the standard comment text including the site address in a case where the mental state is different between the standard comment text and one or more previous comment texts, and/or that instructs the score memory to increase the score of the mental state of one or more subsequent comment texts in a case where the mental state is different between the standard comment text including the site address and one or more subsequent comment texts. The program enables identification of the mental state transition caused by the promotional site address.

Description

Programs, devices for analyzing the effect of promotional sites on the transformation of the user's psychological state and method

technical field

The present invention relates to a technique for analyzing comment text posted to a communication site server such as an SNS (Social Networking Service) site server.

This application claims the benefit of priority under the Paris Convention from Japanese Patent Application No. 2014-145050 filed on July 15, 2014, which is hereby incorporated by reference under PCT Rule 20.6.

Background technique

In recent years, many unspecified people have actively posted their own comments (text information) through the SNS site server. The SNS site server discloses the comment text cast by the users of the group to many other users in the group. Typical examples of SNS are Facebook (registered trademark), Twitter (registered trademark), Google+ (registered trademark), and Mixi (registered trademark), which are also called mini-blog sites.

Review text transmitted from a user often reflects that user's psychological state in a conscious or unconscious manner. Specifically, in marketing, it is preferable to analyze such a comment reflecting the psychological state of the user generated by the product or service. A specific method for analyzing user comments is: obtaining comment texts made by users and profile information of users, and then analyzing content trends of many comment texts according to each profile item. This method is based on the fact that the content of mentioned review texts about products etc. may differ according to user profile information about eg generation, gender and interests.

As an example associated with the above-mentioned techniques, Patent Document 1 discloses a technique of analyzing the effect of a banner advertisement on a web page by using a user's online behavior log. As a behavior log, it is possible to take, for example, the number of clicks on banner advertisements or transitions to purchase behavior performed after clicking on the advertiser's site. The advertiser sets the link destination in order to identify the banner advertisement provided on the link source web page, thereby identifying which banner advertisement is effective.

Also, for example, Patent Document 2 discloses a technique of collecting posted articles mentioning entertainment content, and then retrieving reputation information on the entertainment content. Reputation information in this technique is defined as a numerical value evaluating the positive/negative impression given by an advertisement.

Also, for example, Patent Document 3 discloses a technique of analyzing advertising effects using an AIDMA model, which is an economic model of purchasing behavior. Common sense in economics: The user's mental state related to purchasing behavior has a tendency to change in the order of A (attention), I (interest), D (expectation), M (memory) and A (action). In the technique described in Patent Document 3, one stage of the AIDMA model is assigned to each type of advertisement/behavior (such as TV advertisement, magazine advertisement, email, web page, click on banner, bookmark, get coupon, etc.), and then Measure their number (number of TV ads, number of banners clicked, etc.), which quantifies the impact of advertising on each stage of the AIDMA model.

Also, for example, Patent Document 4 discloses a technique of collecting past contributions of a specific user in SNS, and then estimating the user's profile information (for example, regarding generation and gender) based on features of words appearing in these contributions. The technology can estimate profile information of any user.

reference list

patent documents

Patent Document 1: Japanese Patent Laid-Open Publication No. 2012-088994

Patent Document 2: Japanese Patent No. 4359787

Patent Document 3: Japanese Patent Laid-Open Publication No. 2003-044738

Patent Document 4: Japanese Patent Laid-Open Publication No. 2013-196070

Contents of the invention

The problem to be solved by the present invention

Even in the case where a news article or SNS comment text induces a visit to a purchase site, the technique described in Patent Document 1 has difficulty identifying a web page including a news article or comment text. That is, it was impossible to identify the relationship between the webpage delivered by the third person and the actual store's site, making quantification of the impact on the online promotional site difficult.

Furthermore, the reputation of the promotional site determined by the technique described in Patent Document 2 does not have to be directly linked to the user's purchasing behavior. For example, even a promotional site that gives a positive impression cannot cause a user to purchase a product unless the site provides an adequate and specific description of the product. Conversely, even a promotional site that gives a negative impression may induce purchase of a product if the site is impressive to the user.

Furthermore, the technique described in Patent Document 3 does not assume a technique of estimating the mental state of the AIDMA model based on comment texts posted to SNS. Also, the technique described in Patent Document 4 does not assume temporal changes in profile information.

Facing such a situation, the inventors of the present invention considered the possibility of finding a promotional site having a contribution to a transition of a user's mental state based on a review text about a product or service thrown to an SNS. Furthermore, the possibility of accumulating for each promotional site a score representing the degree of contribution to the transition of the user's mental state with respect to the product or service has been considered. It is assumed that these scores enable a determination of which type of promotional site is effective in marketing.

Therefore, an object of the present invention is to provide a program, apparatus and method for analyzing the effect of promotional sites on user's psychological state transition using review text transmitted by the user.

way of solving the problem

According to the present invention, there is provided a program executed by a computer installed on a device that analyzes the effect of a promotional site for a product or service with respect to a target keyword based on a set of review texts transmitted by each contributor , the effect works on a shift in the user's mental state, the program causes the computer to act as:

A psychological keyword dictionary, through which one or more psychological keywords are registered in advance for each mental state;

a review text obtainer, the review text obtainer obtains review texts for each contributor, each review text includes the target keyword and is arranged in time series;

A mental state determiner, in the case that the obtained comment text of the contributor includes a psychological keyword registered in the psychological keyword dictionary, the mental state determiner outputs the mental state for which the included psychological keyword is registered;

a score memory that stores, for each promotional site address, a score set for each psychological state; and

a score updater that instructs the score store to increase the criteria if the state of mind differs between the standard review text that includes the site address and one or more previous review texts that were generated before the standard review text This rating of the mental state of the review text, and/or, where the mental state differs between the standard review text including the site address and one or more subsequent review texts generated after the standard review text, the The score updater instructs the score store to increase the score of the mental state of the subsequent review text(s).

The program is characterized by enabling the identification of the mental state transition caused by the promotional site address.

As an embodiment of a program executed by a computer according to the present invention, preferably, the psychological keyword dictionary sets these mental states according to the AIDMA (attention, interest, expectation, memory, action) model, and maintains the same relationship with each mental state. One or more registered psychographic keywords corresponding to the target keyword.

As another embodiment of the program executed by the computer according to the present invention, it is also preferred that:

Assign each mental state increasing values in the order A, I, D, M, and A;

Time-series mental state transitions are set to these time-series mental state values; and

The score updater sets each of these time-series values to a moving average derived over a predetermined range.

As another embodiment of the program executed by the computer according to the present invention, it is also preferable that the rating updater set the weight wb of the mental state of the standard comment text and the weight wa of the mental state of the subsequent comment text are different values from each other, and the score is increased by using the weight.

As another embodiment of the program executed by the computer according to the present invention, it is also preferred that the score updater sets the weight wb of the mental state of the standard comment text and the weight wa of the mental state of the subsequent comment text, The weight wb is set to a different value for each mental state, and the weight wa is set to a different value for each mental state, and the score is increased by using the weight.

As another embodiment of the program executed by the computer according to the present invention, it is also preferred that:

The weight wb of each mental state in the mental state AIDM (attention, interest, expectation, memory) is set in advance to a value greater than the weight wb of the mental state A' (action); and

The weight wa of each mental state in the mental state AIDM (attention, interest, expectation, memory) is set in advance to a value smaller than the weight wa of the mental state A' (action).

As another embodiment of the computer-executed program according to the present invention, it is also preferable that, in the case of a different mental state between the standard comment and one or more subsequent comment texts generated after the standard comment, As the posting time interval between the standard review text and the subsequent review text becomes longer, the score updater sets a smaller weight wa.

As another embodiment of the program executed by the computer according to the present invention, it is also preferred that:

The target keyword is a keyword based on the product or service, and

The site address is the URL (uniform resource locator) of the promotional site for the product or service.

As another embodiment of the program executed by the computer according to the present invention, it is also preferable that the program further makes the computer act as a promotional effect notifier, and the promotional effect notifier sends the information stored in the scoring memory for each site address. ratings for each mental state.

According to the present invention, there is also provided a device for analyzing the effect of a promotional site about a product or service of a target keyword based on a set of review texts transmitted by each contributor, which effect has a role in the transition of the user's psychological state. role, and the device includes:

A psychological keyword dictionary, through which one or more psychological keywords are registered in advance for each mental state;

a review text obtainer, the review text obtainer obtains review texts for each contributor, each review text includes the target keyword and is arranged in time series;

A mental state determiner, in the case that the obtained comment text of the contributor includes a psychological keyword registered in the psychological keyword dictionary, the mental state determiner outputs the mental state for which the included psychological keyword is registered;

a score memory that stores, for each promotional site address, a score set for each psychological state; and

a score updater that instructs the score store to increase the criteria if the state of mind differs between the standard review text that includes the site address and one or more previous review texts that were generated before the standard review text This rating of the mental state of the review text, and/or, where the mental state differs between the standard review text including the site address and one or more subsequent review texts generated after the standard review text, the The score updater instructs the score store to increase the score of the mental state of the subsequent review text(s).

The device is characterized by enabling identification of the mental state shift caused by the promotional site address.

According to the present invention, there is further provided a method performed by using a device that analyzes the effect of a promotional site for a product or service with respect to a target keyword based on a set of review texts transmitted by each contributor, the effect Work on the change of the user's psychological state,

The unit includes:

a mental keyword dictionary by which one or more mental keywords are registered in advance for each mental state; and

a score memory that stores, for each promotional site address, a score set for each mental state, and

The method includes the following steps:

Obtain comment texts for each contributor, each comment text includes the target keyword and is arranged in time sequence;

In a case where the obtained review text of the contributor includes a mental keyword registered to the mental keyword dictionary, outputting a mental state for which the included mental keyword is registered; and

instructing the scoring store to increase the score of the mental state of the standard comment text if the mental state differs between the standard review text comprising the site address and one or more previous review texts generated prior to the standard review text , and/or, instructing the scoring store to increase the subsequent(s) if the state of mind differs between the standard comment text comprising the site address and one or more subsequent comment texts generated after the standard comment text This score for this mental state of the review text.

The method is characterized by enabling identification of the mental state shift caused by the promotional site address.

Advantageous Effects of the Invention

The program, device and method according to the present invention allow the use of review texts transmitted by users to analyze the effect of promotional sites on the transformation of the user's psychological state.

Description of drawings

Renders the attached figure, in which:

Figure 1 is a schematic diagram illustrating an embodiment of a system comprising a device according to the invention;

Fig. 2 is a functional block diagram showing an embodiment of the psychological transition analysis device according to the present invention;

3A, 3B and 3C are diagrams for explaining examples of review texts;

Fig. 4 is a schematic diagram for illustrating the first score update process in the present invention;

Fig. 5 is a schematic diagram for illustrating the second score update process in the present invention;

Fig. 6 is a schematic diagram for explaining the first score update process using weights in the present invention, and

FIG. 7 is a schematic diagram for explaining the second score updating process using weights in the present invention.

Detailed ways

Illustrative embodiments of the present invention will now be described with reference to the accompanying drawings.

Figure 1 is a schematic diagram illustrating an embodiment of a system comprising a device according to the invention.

As shown in FIG. 1 , connected to the Internet is a psychotransition analysis device 1 . Psychological transition analysis device 1 is configured to communicate with SNS site server 2 via the Internet. The device 1 can obtain the comment text transmitted by each contributor from the SNS site server 2 through an API (Application Programming Interface). The API is a protocol interface adapted to utilize functions of application services prepared according to server types.

The psychological transition analysis device 1 can accumulate comment texts in the database in advance. That is, the device 1 can choose to communicate with the SNS site server 2 .

An unspecified third person can use his/her own terminal 3 to send the comment text to the SNS site server 2 via the Internet. The psychological transition analysis apparatus 1 is configured to analyze the contributor's psychological state transition caused by a promotional site related to a product or service about a target keyword, using a set of review texts transmitted by each contributor.

The target keyword is a keyword about a product or service. And, among the set of review texts, mixed are review texts including addresses of promotional sites about products or services. The promotional site address is the URL of the promotional site.

(Step S11 ) For example, the user sees a new site by using the terminal 3 . Then, the user transmits to the SNS site server 2 the following comment text including the URL of the new site.

"A new item announcement of the smartphone of α company. http://news.html"

The psychological transition analysis device 1 obtains the above comment texts cast by users. Device 1 analyzes the psychological state reflected in the comment text about "α company's smartphone", thereby judging that the user's psychological state has changed to a mental state that only pays a little attention to α company's smartphone.

(Step S12) After that, the user sees, for example, a product introduction site, and then sends to the SNS site server 2 the following comment text including the URL of the product introduction site.

"I saw Alpha's smartphone on HP. http://hp.html"

The psychological transition analysis device 1 obtains the above comment texts cast by users. The device 1 analyzes the psychological state reflected in the comment text about "α company's smartphone", thereby judging that the user's psychological state has changed to a psychological state that the user is "interested" in the α company's smartphone.

As described above, when the user's mental state changes from the mental state before seeing the site after viewing the promotional site, the device 1 judges that the promotional site contributes to the transition of the mental state. Accumulation of judgment results makes it possible to estimate the effect of the promotional site on the transition of the psychological state.

Fig. 2 is a functional block diagram showing an embodiment of the psychotransition analysis device according to the present invention.

The psychological transition analyzing device 1 according to the present invention is configured to analyze the psychological state transition caused by the promotion site about the target keyword using the review text of each contributor. The device 1 is configured to include a psychological keyword dictionary 101 , a score storage 102 , a review text obtainer 11 , a mental state determiner 12 , a score updater 13 , and a promotional effect notifier 14 in addition to a communication interface. These functional units are embodied by executing corresponding programs on a computer installed on the device, in addition to the communication interface. In addition, the processes in these functional units can be understood as a method for analyzing the psychotransformation performed in the device.

(Psychological Keyword Dictionary 101 ) The psychological keyword dictionary 101 is configured to record one or more psychological keywords for each psychological state in advance. According to the AIDMA model, psychological states are considered to change with the following time series.

Cognitive stage: attention (A) -->

Emotional stage: interest (I) -->

Expectation (D)——>

Memory (M) -->

Action phase: action (A) -->

The AIDMA model is a model for describing the psychological process of consumers to advertisements. This purchasing psychology transition economic model enables the estimation of the psychological process by tracking specific user opinions about a target keyword such as a target product or service in time series.

The psychological keyword dictionary 101 records one or more psychological keywords regarding keywords such as "smartphone of α company" for each of the psychological states (A, I, D, M, and A). E.g,

Follow status: launch, commercial, new project, new product, ...;

Interest status: product line, manual, HP, friend, ...;

Desired state: want, price, specification, design, ...;

Memory state: store, display, actual product, preview, ...; and

Action status: buy, get, give up, model change, ...

The registration is performed in the form of a dictionary including characteristic words about the corresponding psychological state of the target keyword.

Note that the mental keyword dictionary 101 is usually prepared manually. However, by using manually recorded psychological keywords as seeds (teacher data), statistical indicators such as X ² (chi-square) values or Akeck Information Criterion (AIC) can be used to characterize the corresponding psychological states words.

(Review Text Obtainer 11) The Review Text Obtainer 11 is configured to time by time for each contributor (each user ID) by using a target keyword (for example, "a company's smartphone") as a retrieval key The sequence obtains comment text from the SNS site server 2 . The target keyword can be the name of a product or service. The obtained review text is output to the mental state determiner 12 .

3A, 3B, and 3C are diagrams for explaining examples of comment text.

As shown in FIG. 3A , various comment texts on the target keyword (smartphone of α company) are obtained for many users. Among review texts, mixed are review texts including promotional site addresses (URLs) about products or services. According to the invention, the included address of the promotional site makes it possible to estimate the effect of the promotional site on the transformation of the user's psychological state.

According to FIG. 3B , the review text about the target keyword (smartphone of α company) is narrowed down to the review text about the target keyword for user AAA.

(Mental state determiner 12) Returning to Fig. 2, the mental state determiner 12 is configured to output and include the included psychological keywords when the review text of each contributor includes the psychological keywords registered to the psychological keyword dictionary. corresponding mental state. For example, by using review text including "smartphone of α company" as a target keyword and "new item" as a psychological keyword, the mental state determiner 12 can estimate that the user is in an attention (A) state. And, based on the review text including "smartphone of α company" as the target keyword and "commercial advertisement" as the psychological keyword, it can be estimated that the user is in the attention (A) state. Further, it can be estimated that the user is in the state of interest (I) based on the review text including "smartphone of α company" as the target keyword and "HP" as the psychological keyword. Furthermore, by using review text including "smartphone of α company" as a target keyword and "want" as a psychological keyword, the mental state determiner 12 can estimate that the user is in a desired (D) state. In addition, it can be estimated that the user is in the memory (M) state from review texts including "smartphone of α company" as a target keyword and "store" as a psychological keyword.

As shown in FIG. 3C , the mental state is estimated for each comment text of the user AAA about the target keyword (smartphone of α company).

Alternatively, in a case where a plurality of psychological keywords are included in one review text, the psychological state corresponding to the entire review text may be determined based on the frequency of occurrence of each of the plurality of psychological keywords. Further, discriminators such as support vector machines can be used to determine mental states based on the occurrence trends of words in the teacher data. Furthermore, psychological states can be analyzed by using likelihood as well as AIDMA models.

Regarding "New Project Release of α Company's Smartphone";

"NewItem" = A, "Release" = A,

A=2, I=0, D=0, M=0, A=0

Therefore, the probability of A = 100% (one hundred percent)

(Score Memory 102) The score memory 102 is configured to store a score of each mental state for each site address.

(Score Updater 13) The score updater 13 is configured to perform two main steps: "mental shift determination step" and "score update step".

(Psychological Transition Determination Step) The score updater 13 is adapted to estimate the psychological transition, which becomes data in which the mental states are arranged in time series for each contributor. For example, the following psychological shifts are estimated;

A, A, I, A, I, I, D, I, D, D, M, M, M

Here, each mental transition is preferably assigned a series of numerical values, each numerical value in the series of numerical values is in the order of A, I, D, M, A (A->I->D->M->A) get bigger. This numerical assignment represents psychological transitions as a time series of numerical values, where each numerical value corresponds to a psychological state. The quantity of values in a time series can be understood as an index indicating whether a transition to the next mental state occurs.

A=0, I=1, D=2, M=3

A, A, I, A, I, I, D, I, D, D, M, M, M

0,0,1,0,1,1,2,1,2,2,3,3,3

In such a psychological transformation, the mental states will naturally occur in sequence.

<Correction of Psychological Shift by Moving Average> The score updater 13 is adapted to correct the numerical values arranged in time series as completely as possible to the sequence of these values in ascending order. Specifically, preferably, the mental state of the target comment is determined as the average value of the mental states of n adjacent comments. Alternatively, the mental state of the target review may be a moving average of previous reviews within a predetermined range. The following is an example in the case of averaging two reviews adjacent to each other.

A, A, I, A, I, I, D, I, D, D, M, M, M

0,0,1,0,1,1,2,1,2,2,3,3,3

0.3, 0.3, 0.4, 0.6, 1.1, 1.4, 1.6, 2, 2.2, 2.4, 2.6, 2.75, 3 rounded to integer ->

0,0,0,1,1,1,1,2,2,2,2,3,3,3

A, A, A, I, I, I, I, D, D, D, D, M, M, M

<Correction for Mental Transition Based on Transition Time Difference>Commentary text usually includes transit time. Then, the mental state determiner 12 is adapted to output the transmission time of the review text including the psychological keywords together with the mental state of the review text. The score updater 13 may use adjacent two transmission times to record the elapsed time interval between mental states arranged in time series.

The score updater 13 then weights each of the chronologically arranged values in such a way that:

The greater the previous elapsed time interval, the smaller the weighting factor used to weight the values; and

The shorter the previous elapsed time interval, the larger the weighting factor used to weight the values.

Alternatively, each of the integers 0-4 corresponding to a mental state may be weighted by the estimated likelihood of the corresponding mental state. In the case of A=10%, I=70%, D=20% and M=0%, the integers 0-3 are weighted as follows:

0*10%+1*70%+2*20%+3*0%＝1.1

(Score Updating Step) The score updater 13 is also adapted to perform either or both of the following two update steps, and then give instructions to the score memory 102 .

<Standard update step> A standard update step is a step in which, in a set of review texts, when the standard review text and one or more previous review texts defined as review texts generated before the standard review text When the mental state is different, increase the mental state score of the standard comment text that includes the site address.

<Subsequent update step> The subsequent update step is a step in which, when the mental state is different between the standard review text and the subsequent review text, one or more subsequent updates defined as the review text generated after the standard review text are added. Mental state ratings of review texts.

Specifically, the AIDMA transitions in the previous/subsequent review text before/after the standard review text during these steps are expressed as follows:

Previous Comment Text --> Standard Comment Text:

db(a, i, d, m, a')

Standard Comment Text --> Subsequent Comment Text:

da(a, i, d, m, a')

Here, each component with (a, i, d, m, a') has a value of 0 or 1.

Fig. 4 is a schematic diagram for explaining the first score updating process in the present invention.

(step S41) the mental state of user AAA changes as follows:

"I happened to see many commercials (CMs) for Alpha's smartphones." (Text of previous comment) -->Follow A

(Browse site: http://hp.html)

"I saw a smartphone from α company on HP. http://hp.html" (standard comment text) --> Interest I

In this way, by viewing the site: http://hp.html, user AAA's mental state changes from Concern A to Interest I. Therefore, the site http://hp.html was judged as contributing to the change of mental state to interest I. When such a judgment is made, the score P(I) of the interest I for the site http://hp.html is incremented by 1 in the score memory 102 .

Previous Comment Text --> Standard Comment Text:

db(0,1,0,0,0)

(Step S42) Then, the mental state of the user AAA changes as follows:

"I saw a smartphone from α company on HP. http://hp.html" (standard comment text) --> Interest I

"I want a smartphone from company α. RT@XXX" (subsequent comment text) --> Expectation D

In this way, user AAA's state of mind changes from Interest I to Desire D by viewing the site: http://hp.html. Therefore, the site http://hp.html was judged to have contributed to the change of mental state to Desired D. When such a judgment is made, the score P(D) of the expectation D for the site http://hp.html is incremented by 1 in the score memory 102 .

Standard Comment Text --> Subsequent Comment Text:

da(0,0,1,0,0)

By accumulating db(a, i, d, m, a') and da(a, i, d, m, a') as explained above, the score of each promotional site can finally be estimated. In scoring memory 102 as shown in Figure 2, the following estimates are made:

News site page A (http://news.html) has a greater contribution to the mental state “Attention (A)”.

Product introduction site page B (http://hp.html) has the largest contribution to the mental state "Action (A')" and the second largest contribution to the mental state "Expectation (D)".

Subject site page C (http://aaa.html) has the largest contribution to the mental state "Attention (A)" and the second largest contribution to the mental state "Interest (I)" and the mental state "Expectation (D)" .

Fig. 5 is a schematic diagram for explaining the second score updating process in the present invention.

(step S51) the mental state of user AAA changes as follows:

"The new project of α company's smart phone is released. http://news.html"-->Follow A

"The commercial for Alpha's smartphone is interesting." -->Interest I

In this way, by viewing the site: http://news.html, the mental state of user AAA changes from concern A to interest I. Therefore, the site http://news.html was judged as contributing to the change of state of mind to Interest I. When such a judgment is made, the score P(I) of the interest I for the site http://news.html is incremented by 1 in the score memory 102 .

(step S52) Then, the mental state transition of user AAA is as follows:

"The commercial for Alpha's smartphone is interesting." -->Interest I

"I saw the α company's smartphone on HP. http://hp.html" --> Interest I

At this time, even by viewing the site: http://hp.html, the mental state of the user AAA does not change and maintains the state of interest I. Therefore, no score is updated in the score memory 102 .

(Step S53) Then, the mental state of user AAA changes as follows:

"I saw the α company's smartphone on HP. http://hp.html" --> Interest I

"My friend has a smartphone from company α" --> Interest I

Again, even by viewing the site: http://hp.html, user AAA's mental state did not change and maintained the state of interest I. Therefore, no score is updated in the score memory 102 .

FIG. 6 is a schematic diagram for explaining the first score update process using weights in the present invention.

As shown in FIG. 6 , a score update process similar to that shown in FIG. 4 is performed. As will be seen below, the score update process shown in Figure 6 differs in that the score for each mental state is increased by using different weights in each mental state.

(S61) Weight wb (previous weight) for the mental state of the standard comment text:

wb(A), wb(I), wb(D), wb(M), wb(A')

(S62) For the weight wa of the mental state of the subsequent comment text (the weight afterwards), the weight wa is not equal to wb:

wa(A), wa(I), wa(D), wa(M), wa(A')

In the above cases, the scoring value P is expressed as follows:

P＝Σi _＝A-A' (wb(i)*db(i)+wa(i)*da(i))

Set the weight wb of each of mental states A, I, D, and M (attention-->interest-->expectation-->memory) to a value greater than the weight wb of mental state A' (action) . In increasing the scoring of mental states in the standard review text, the above setting means that promotional sites described in standard review texts where the mental state changes to A, I, D, or M contribute more to the shift in mental state than in the psychological Contribution of promotional sites to mental state transitions described in the standard review text for state changes to A'.

Further, the weight wa of each mental state in mental states A, I, D and M (attention-->interest-->expectation-->memory) is set to be less than the weight of mental state A' (action) The value of wa. In increasing the scoring of mental states in subsequent review texts, the above settings imply that promotional sites described in standard review texts with a change in mental state to A, I, D, or M contribute less to the shift in mental state than in psychological Contribution of promotional sites to mental state transitions described in the standard review text for state changes to A'.

FIG. 7 is a schematic diagram for explaining the second score updating process using weights in the present invention.

As shown in Figure 7, when the mental state between the standard comment text and one or more subsequent comment texts generated after the standard comment text is different, with the posting time interval between the standard comment text and the subsequent comment text becomes longer, the score updater 13 sets the weight wa to a smaller value. At this time, not only one comment text immediately before/after the standard comment text, but also a plurality of comment texts located near the standard comment text are considered. Then, set the weights as follows:

A greater weight is set for mental states generated from review texts that are temporally or positioned closer to the standard review text.

Smaller weights are set for mental states generated by review texts that are temporally or located further away from the standard review text.

(Promotional Effect Notifier 14) The promotional effect notifier 14 is configured to transmit the score of each mental state stored in the score memory 102 for each site address.

As explained in detail above, the program, apparatus and method according to the present invention allow the use of review texts transmitted by users to analyze the effect of promotional sites on transitions based on the user's psychological state. At this time, the more effective the promotion site is to the user's purchase behavior, the more valuable the promotion site is. The present invention not only refers to the psychological state included in the review text, but also enables the estimation of the effect of the promotional site based on the change of the user's psychological state with regard to the purchase. Specifically, online promotional effects of news articles or SNS comments (of which it is generally impossible to estimate promotional effects) can be estimated.

Many widely different alternatives and modifications of the above-described various embodiments of the invention may be constructed without departing from the spirit and scope of the invention. All of the foregoing examples are merely exemplary of the invention and are not intended to be limiting. Accordingly, the invention is to be limited only as defined in the appended claims and their equivalents

List of reference signs

1 Psychological transition analysis device; 101 Psychological keyword dictionary; 102 Score memory; 11 Review text obtainer; 12 Mental state determiner; 13 Score updater; 14 Promotion effect notifier; 2 SNS site server; 3 Terminal; server.

Claims (11)

1. A computer-readable storage medium storing therein a program executed by a computer installed on a device that promotes a product or service about a target keyword based on a set of review texts transmitted by each contributor Analyze the effect of the site, which is caused by the change of the user's psychological state, The device has: A psychological keyword dictionary, through which one or more psychological keywords are registered in advance for each mental state; a scoring memory storing, for each promotional site address, a scoring set for each mental state, When the program is executed by a computer, the following steps are realized: As a comment text obtainer, the step of obtaining comment text for each contributor, each comment text includes the target keyword and is arranged in time series; As a mental state determiner, in the case that the obtained comment text of the contributor includes the psychological keywords registered in the psychological keyword dictionary, outputting the step of registering the mental state targeted by the included psychological keywords; As a scoring updater, instructing the scoring memory to increase the value of the standard commenting text in the case of a different mental state between the standard commenting text comprising the site address and one or more previous commenting texts generated before the standard commenting text The score of the mental state, and/or, in the case of a different mental state between the standard review text including the site address and one or more subsequent review texts generated after the standard review text, indicates the score memory The step of increasing the scoring of the mental state of the one or more subsequent review texts and enabling identification of the mental state shift caused by the promotional site address. 2. The computer-readable storage medium as claimed in claim 1, wherein, the psychological keyword dictionary sets these mental states according to the AIDMA (attention, interest, expectation, memory, action) model, and registers with the mental state for every kind of mental state One or more psychographic keywords corresponding to the target keyword. 3. The computer readable storage medium of claim 2, where each mental state is assigned a value that increases in the order of mental state concern, mental state interest, mental state expectation, mental state memory, and mental state action, where the time-series mental state transitions are set to these time-series mental state values, and Therein, the program acts as the score updater to set each of these time-series values to a moving average derived over a predetermined range. 4. The computer-readable storage medium as claimed in claim 2 , wherein the program serves as the score updater to set the weight of the mental state of the standard review text and the weight of the mental state of the subsequent review text to be different from each other , and the score is increased by using weights. 5. The computer-readable storage medium as claimed in claim 2 , wherein the program sets the weight of the mental state of the standard comment text and the weight of the mental state of the subsequent comment text as the score updater, for each Mental states set weights to different values and increase the score by using weights. 6. The computer readable storage medium of claim 4, Wherein, the program is used as the score updater to set the weight of each mental state in the mental state attention, mental state interest, mental state expectation, and mental state memory of the standard comment text to be greater than the psychological state for the standard comment text in advance. the value of the weight of the state action, and Wherein, the weight of each mental state in the mental state attention, mental state interest, mental state expectation, and mental state memory for the subsequent comment text is set in advance to a value smaller than the weight of the mental state action for the subsequent comment text. 7. The computer-readable storage medium as claimed in claim 4, wherein, under the condition that the mental state between the standard comment text and one or more subsequent comment texts generated after the standard comment text is different, when the As the posting time interval between the standard review text and the subsequent review text becomes longer, the program as the score updater sets a smaller weight for the mental state of the subsequent review text. 8. The computer readable storage medium of claim 1, where the target keyword is a keyword based on the product or service, and Wherein, the site address is the URL (uniform resource locator) of the promotional site of the product or service. 9. The computer-readable storage medium as claimed in claim 1, when the program is executed by a computer, it further implements the step of sending the scoring of each mental state stored in the scoring memory for each site address as a promotion effect notifier . 10. A device for analyzing the effect of a promotional site about a product or service of a target keyword, which is caused by a user's mental state transition, based on a set of review texts transmitted by each contributor, and the device comprises : A psychological keyword dictionary, through which one or more psychological keywords are registered in advance for each mental state; a review text obtainer, the review text obtainer obtains review texts for each contributor, each review text includes the target keyword and is arranged in time series; A mental state determiner, in the case that the obtained comment text of the contributor includes a psychological keyword registered in the psychological keyword dictionary, the mental state determiner outputs the mental state for which the included psychological keyword is registered; a score memory that stores, for each promotional site address, a score set for each psychological state; and a score updater that instructs the score store to increase the criteria if the state of mind differs between the standard review text that includes the site address and one or more previous review texts that were generated before the standard review text This rating of the mental state of the review text, and/or, where the mental state differs between the standard review text including the site address and one or more subsequent review texts generated after the standard review text, the Score updater instructs the score store to increase the score of the mental state of the subsequent comment text, and The means enable identification of the mental state shift caused by the promotional site address. 11. A method of analyzing the effect of a promotional site about a product or service of a target keyword caused by a user's psychological state transition based on a group of review texts transmitted by each contributor by using a device , The unit includes: a mental keyword dictionary by which one or more mental keywords are registered in advance for each mental state; and a score memory that stores, for each promotional site address, a score set for each mental state, and The method includes the following steps: Obtain comment texts for each contributor, each comment text includes the target keyword and is arranged in time sequence; In a case where the obtained review text of the contributor includes a mental keyword registered to the mental keyword dictionary, outputting a mental state for which the included mental keyword is registered; and instructing the scoring store to increase the score of the mental state of the standard comment text if the mental state differs between the standard review text comprising the site address and one or more previous review texts generated prior to the standard review text , and/or instructing the scoring memory to increase the psychometric of the subsequent comment text if the psychological state of the standard comment text comprising the site address is different from one or more subsequent comment texts generated after the standard comment text the rating of the state, and The method enables identification of the mental state shift caused by the promotional site address.