JP2020140405A - Information processing system, information processing method, and program - Google Patents

Abstract

To appropriately provide both positive word-of-mouth and negative word-of-mouth to a user.SOLUTION: An information processing system of the present invention is an information processing system that extracts first word-of-mouth information included in advertisement information related to an advertising product for each user, and that is provided with a word-of-mouth extraction unit for extracting word-of-mouth information including a negative description about a first function in the word-of-mouth information from the user for the purchased product of the user, from the word-of-mouth information for the advertising product as first word-of-mouth information to be included in the advertisement information.SELECTED DRAWING: Figure 1

Description

The present invention relates to information processing systems, information processing methods, and programs.

Patent Document 1 describes an information processing device that provides a mechanism for utilizing word-of-mouth information for sales promotion. The information processing device described in Patent Document 1 is provided to the user using a score calculated based on the user's evaluation information for each feature item representing the attributes of the product or service such as function, effect, and utility. Word-of-mouth information is selected from the word-of-mouth information storage. Here, the user's evaluation information includes the degree of importance, which is the degree to which the user attaches importance to the characteristic item, and the degree of attractiveness, which is the degree to which the user is attracted to the explanatory information corresponding to the characteristic item of the product. In addition, the score is calculated based on the importance, attractiveness, and experience. Then, in the information processing apparatus described in Patent Document 1, word-of-mouth information including a feature item having a high score is selected as information for sales promotion. When the word-of-mouth information is selected from the word-of-mouth information storage unit, for example, non-positive word-of-mouth information and negative word-of-mouth information are excluded.

Japanese Unexamined Patent Publication No. 2011-86056

By the way, for example, the following two methods can be considered for determining whether or not to purchase a product or service (hereinafter, represented by a product) with reference to word-of-mouth information. One is when a plurality of word-of-mouth information are all positive. The other is when multiple word-of-mouth information contains both positive and negative ones. After actually purchasing the product, when it is good to purchase it, in the case of the former, it seems that there are some purchasers who feel that it was correct to make a decision to purchase by referring to word-of-mouth information, for example. On the other hand, in the latter case, in addition to feeling that the decision to purchase with reference to word-of-mouth information was correct, the decision to purchase after comprehensively considering positive word-of-mouth information and negative word-of-mouth information. It seems that some buyers feel that they were right. In these cases, it seems that there are not a few purchasers who feel that the degree of satisfaction (or the degree of affirmation) that the judgment was correct is higher in the latter purchase decision than in the former purchase decision. Is done.

On the other hand, from the viewpoint of the seller of the product, the negative word-of-mouth information is the information in the direction of suppressing the sales, and is the information that the purchaser does not want to provide as much as possible.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide an information processing system, an information processing method, and a program capable of appropriately providing both positive reviews and negative reviews to users. And.

In order to solve the above problems, the information processing system of the present invention is an information processing system that extracts the first word-of-mouth information included in the advertising information related to the advertising product for each user, and the information processing system for the purchased product of the user. It is characterized by including a word-of-mouth extraction unit that extracts word-of-mouth information including a negative description about the first function in the user's word-of-mouth information from the word-of-mouth information for the advertisement product as the first word-of-mouth information to be included in the advertisement information. ..

Further, in the information processing system of the present invention, the word-of-mouth extraction unit is word-of-mouth information including a negative description about the first function of the advertising product from a plurality of word-of-mouth information related to the advertising product. It is characterized in that the first score according to the evaluation of the advertising product as a whole is within a predetermined range with respect to the second score which is an index of the evaluation of the advertising product, as the first word-of-mouth information. And.

Further, in the information processing system of the present invention, when the word-of-mouth extraction unit includes a positive description about the second function of the purchased product in the word-of-mouth information related to the purchased product of the user, a plurality of the word-of-mouth extraction unit related to the advertising product. It is characterized in that the word-of-mouth information including a positive description about the second function of the advertising product is extracted as the second word-of-mouth information to be included in the advertising information together with the first word-of-mouth information.

Further, in the information processing system of the present invention, when the word-of-mouth extraction unit includes a negative description about the third function of the browsed product, the word-of-mouth information relating to the browsed product that the user has browsed but has not purchased is included. It is characterized in that it does not extract information containing a negative description about the third function of the advertising product from a plurality of word-of-mouth information related to the advertising product.

Further, the information processing system of the present invention is further characterized by further including an embedded unit that incorporates the first word-of-mouth information extracted by the word-of-mouth extraction unit into print information for a delivery printed matter of an envelope, a sealed letter, a postcard, or a home delivery. To do.

Further, the information processing method of the present invention is an information processing method for extracting word-of-mouth information included in the advertising information related to the advertising product for each user, and the word-of-mouth extraction unit uses the user's purchased product for the user. It is characterized by including a process of extracting word-of-mouth information including a negative description about the first function in word-of-mouth information from word-of-mouth information for the advertisement product as the first word-of-mouth information to be included in the advertisement information.

Further, the information processing method of the present invention is further characterized by further including a process of incorporating the first word-of-mouth information extracted by the word-of-mouth extraction unit into print information for a delivery printed matter of an envelope, a sealed letter, a postcard or a home delivery. ..

Further, the program of the present invention is a program that causes a computer to execute a function of an information processing system that extracts word-of-mouth information included in advertising information related to an advertising product for each user, and the computer is used as a purchased product of the user. To function as a word-of-mouth extraction means for extracting word-of-mouth information including a negative description about the first function in the word-of-mouth information of the user as the first word-of-mouth information included in the advertisement information. It is a program.

In addition, the program of the present invention causes the computer to function as a means for incorporating the first word-of-mouth information extracted by the word-of-mouth extraction unit into print information for an envelope, a sealed letter, a postcard, or a delivery printed matter of a home delivery. It is a program of.

According to each aspect of the present invention, both positive word-of-mouth and negative word-of-mouth can be appropriately provided to the user.

It is a block diagram which shows the structural example of the information processing system which concerns on one Embodiment of this invention. It is a schematic diagram which shows the structural example of the product evaluation DB 31 shown in FIG. It is a schematic diagram which shows the structural example of the user attribute DB 32 shown in FIG. It is a schematic diagram which shows the structural example of the word-of-mouth evaluation DB 33 shown in FIG. It is a schematic diagram which shows the structural example of the word-of-mouth evaluation DB 33 shown in FIG. It is a schematic diagram which shows the operation example of the information processing system 1 shown in FIG. It is a flowchart which shows the operation example of the product analysis part 21 shown in FIG. It is a flowchart which shows the operation example of the user analysis unit 22 shown in FIG. It is a flowchart which shows the operation example of the word-of-mouth extraction part 23 and the built-in part 24 shown in FIG. It is a schematic diagram which shows the operation example of the product analysis part 21 shown in FIG. It is a schematic diagram which shows the operation example of the user analysis unit 22 shown in FIG. It is a schematic diagram which shows the operation example of the user analysis unit 22 shown in FIG. It is a schematic diagram which shows the operation example of the word-of-mouth extraction unit 23 shown in FIG. It is a schematic diagram which shows the operation example of the word-of-mouth extraction unit 23 shown in FIG. It is a schematic diagram which shows the operation example of the information processing system 1 shown in FIG.

Hereinafter, an information processing system according to an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration example of an information processing system 1 according to an embodiment of the present invention. The information processing system 1 shown in FIG. 1 is a system composed of one or a plurality of computers such as a server and a personal computer, and each computer includes a CPU (central processing unit), a storage device, an input / output device, and the like, and is included in the storage device. It operates by executing a predetermined program that is stored, and for example, extracts word-of-mouth information included in the advertisement information related to the advertisement product for each user. In the following, a case where cosmetics are used as advertising products will be described as an example. Further, the purchase history, browsing history, word-of-mouth information, etc., which will be described later, may be stored and managed in the information processing system 1, and other computers connected to the information processing system 1 by, for example, a communication line (not shown). It may be stored and managed within.

In the present embodiment, the information processing system 1 includes a processing unit 2 and a storage unit 3 as a functional configuration composed of a combination of hardware and software such as a program. Further, the processing unit 2 includes a product analysis unit 21, a user analysis unit 22, a word-of-mouth extraction unit 23 (word-of-mouth extraction means), and an embedding unit 24 (embedding means). Further, the storage unit 3 includes a product evaluation DB (database) 31, a user attribute DB 32, and a word-of-mouth evaluation DB 33. The product analysis unit 21 analyzes the word-of-mouth information of the product and accumulates the analysis results in the product evaluation DB 31 and the word-of-mouth evaluation DB 33. The user analysis unit 22 analyzes each user based on the purchase history, browsing history, etc. of the product, and accumulates the analysis result in the user attribute DB 32. The word-of-mouth extraction unit 23 refers to the product evaluation DB 31 and the user attribute DB 32, and extracts word-of-mouth information to be included in the advertisement information related to the advertisement product for each user from the word-of-mouth evaluation DB 33. The embedding unit 24 incorporates the word-of-mouth information extracted by the word-of-mouth extraction unit 23 into the print information for the delivery printed matter of the envelope, the envelope, the postcard, or the home delivery.

Here, a configuration example of the product evaluation DB 31 shown in FIG. 1 will be described with reference to FIG. FIG. 2 is a schematic diagram showing a configuration example of the product evaluation DB 31 shown in FIG. The product evaluation DB 31 shown in FIG. 2 is a file for storing information indicating the evaluation result of the product, and includes information about five products in the example shown in FIG. In this case, the product evaluation DB 31 sets each record corresponding to each product as "product ID (identification code)", "category ID", "product name", "introduction", "pores, keratin", "acne," It consists of the fields of "shiny", "wrinkles", "elasticity, sagging", "whitening, spots", "sensitive skin", "moisturizing", "price" and "comprehensive evaluation".

Product identification information is stored in the "product ID" field. In the example shown in FIG. 2, the "product ID" is "1", "2", "3", "4", and "5" in this order from the top. The product category identification information is stored in the "category ID" field. In this embodiment, products of the same or similar genre are classified into the same category. In the example shown in FIG. 2, the "product ID" is "1", the "category ID" of each product of "2" and "3" is "1" (basic cosmetics), and the "product ID" is "4". The "category ID" of each product of "5" is "2" (shampoo, etc.). The "product name" is the product name of the product, and the "introduction text" is the product description of the product. In addition, products are available in the fields of "pores, keratin," "acne, shine," "wrinkles," "elasticity, sagging," "whitening, age spots," "sensitive skin," "moisturizing," and "price." Numerical values representing the sentiment analysis results for each function of are stored. In the present embodiment, each function of the product is an evaluation unit of the product, and includes the role, action, characteristics, features, and the like of the product.

The function of "pores and keratin" is, for example, the action of pore care and keratin care. The function of "acne and shine" is, for example, an action of eliminating acne and shine. The function of "wrinkles" is, for example, an action of improving wrinkles. The function of "elasticity and slack" is, for example, an action of increasing elasticity or decreasing slack. The function of "whitening, spots" is, for example, the action of whitening or the action of reducing spots. The function of "sensitive skin" is, for example, an action of having a small effect on sensitive skin. The function of "moisturizing" is, for example, an action of keeping moisturizing. The function of "price" is a characteristic of whether or not the price is high or whether or not the cost performance is high.

In addition, the sentiment analysis result for each function is, for example, a description related to a product made by a user or the like on an introductory text, a mail-order sales site, a price comparison site, an SNS (social networking service), or the like. Alternatively, it is called word-of-mouth information) can be expressed as a numerical value indicating whether each function of each product is evaluated positively or negatively by analyzing sentences using a sentiment analysis tool or the like. In the present embodiment, the numerical value representing the sentiment analysis result is a real number from "-5" to "5", "5" has the highest degree of positive side (positive), and "0" has no description. Then, "-5" indicates that the degree of negative side (negative) is the highest. In addition, a numerical value representing the overall evaluation of the product is stored in the "comprehensive evaluation" field. In the present embodiment, the numerical value representing "comprehensive evaluation" is a real number from "1" to "5", and "5" has the highest degree of positive side (positive) and "1" has the lowest degree. Represent.

In the present embodiment, the word-of-mouth information is information including one or more sentences described by the user or the like, and the attribute information such as the name, user name, handle name, age, age, etc. of the described user and the description. It may include supplementary information such as the date and time and the name of the area described. Further, the function of the product shown in FIG. 2 is an example, and may include, for example, a function related to odor and scent, a function related to capacity and size, a function related to package design, and the like.

Next, a configuration example of the user attribute DB 32 shown in FIG. 1 will be described with reference to FIG. FIG. 3 is a schematic diagram showing a configuration example of the user attribute DB 32 shown in FIG. The user attribute DB 32 shown in FIG. 3 is a file that stores information representing the user's attribute and the content of recognition of each function related to the user's cosmetics. In the present embodiment, the content of recognition for each function is whether the high or low evaluation of the function is felt as an advantage (value, advantage, advantage, advantage, etc.) or a disadvantage (disadvantage, disadvantage, disadvantage, etc.). Etc.), or neither (neutral) or not (neutral). In the example shown in FIG. 3, the user attribute DB 32 includes information about four users A to D. In this case, the user attribute DB 32 sets each record corresponding to each user as "user ID", "name", "age", "skin quality", "pores, keratin", "acne, shine", "wrinkles". , "Resilience, sagging", "whitening, blemishes", "sensitive skin", "moisturizing" and "price".

User identification information is stored in the "User ID" field. In the example shown in FIG. 3, the "user ID" is "1", "2", "3", and "4" in this order from the top. The "Name", "Age", and "Skin type" fields store a character string representing the user's name, age, and skin type.

In addition, users in the fields of "pores, keratin", "acne, shine", "wrinkles", "elasticity, sagging", "whitening, age spots", "sensitive skin", "moisturizing", and "price" A numerical value representing the content of recognition for each function of the product of is stored. In the present embodiment, the numerical value representing the content of recognition is a real number from "-5" to "5", and "5" is most likely to be a merit, and "0" is neutral (careful). (No), and "-5" indicates that the degree of disadvantage is the highest. Here, feeling a merit means that a high evaluation of the function is a merit for the user. Being neutral means that even if the evaluation of the function is low, the user does not care. And, feeling a disadvantage means that a low evaluation of the function is felt to be a disadvantage by the user. Numerical values representing the content of recognition for each function of the product include, for example, the purchase history of the product of the user, the browsing history (history of the product that was browsed but did not reach the purchase), and the function of each product in the product evaluation DB 31. Based on the content of each evaluation, for example, it can be determined as follows. That is, for purchased products, each function classified as having a positive evaluation content in the product evaluation DB 31 is regarded as a merit function, and each function classified as having a negative evaluation content is neutral. It is a function that does not bother you. For products that have been viewed but not purchased, each function that is classified as having a positive evaluation content in the product evaluation DB 31 is regarded as a merit function, and the evaluation content is classified as a negative side. Each function is a demerit function. Regarding quantification, for example, for a plurality of products, points are added in the positive direction, no points are added, or points are added in the negative direction based on the number of times classified into merit, neutral, or demerit for each function. The degree of merit, neutral, or demerit can be quantified.

However, the numerical value representing the recognition content stored in the user attribute DB 32 may be determined based on, for example, the result of the user responding to a questionnaire or the like.

Next, a configuration example of the word-of-mouth evaluation DB 33 shown in FIG. 1 will be described with reference to FIGS. 4 and 5. 4 and 5 are schematic views showing a configuration example of the word-of-mouth evaluation DB 33 shown in FIG. In the present embodiment, the word-of-mouth evaluation DB 33 shown in FIG. 1 includes the word-of-mouth evaluation DB 33-1 for each word-of-mouth sentence shown in FIG. 4 and the word-of-mouth evaluation DB 33-2 for each word-of-mouth shown in FIG.

The word-of-mouth evaluation DB33-1 for each word-of-mouth sentence shown in FIG. 4 is a file that stores information indicating the result of sentiment analysis for one word-of-mouth sentence written regarding the function. In the example shown in FIG. 4, the word-of-mouth evaluation DB33-1 contains information about six sentences. In this case, the word-of-mouth evaluation DB33-1 sets each record corresponding to each sentence as "word-of-mouth ID", "product ID", "sentence ID", "word-of-mouth sentence", "function ID", "function word" and "function word". It consists of each field of "Negative / Positive Score". The "word-of-mouth ID" is identification information of each word-of-mouth. The "product ID" is the above-mentioned product identification information. The "sentence ID" is identification information of each sentence in one word of mouth. In the example shown in FIG. 4, for example, a word-of-mouth with a “word-of-mouth ID” of “1” includes three sentences with a “sentence ID” of “1” to “3”. A character string representing each sentence is stored in the "word-of-mouth sentence" field. The function identification information included in each sentence is stored in the "function ID" field. In the present embodiment, the "function ID" of the function related to "pores and keratin" is "0", the "function ID" of the function related to "acne and shine" is "1", and the "function" of the function related to "wrinkles". "ID" is "2", "function ID" of the function related to "elasticity, sagging" is "3", "function ID" of the function related to "whitening, spots" is "4", function related to "sensitive skin" The "function ID" of the above is "5", the "function ID" of the function related to "moisture" is "6", and the "function ID" of the function related to "price" is "7". In the "function word" field, a character string indicating a word related to the function included in each sentence is stored. Then, in the field of "negative / positive score", a numerical value representing the above-mentioned sentiment analysis result from "-5" to "5" is stored.

On the other hand, the word-of-mouth evaluation DB 33-2 for each word-of-mouth shown in FIG. 5 is a file that stores information in which the evaluation results of sentiment analysis are averaged and integrated for one word-of-mouth. In the example shown in FIG. 5, the word-of-mouth evaluation DB 33-2 contains information about four word-of-mouth. In this case, the word-of-mouth evaluation DB 33-2 sets each record corresponding to each word-of-mouth as "word-of-mouth ID", "product ID", "pores, keratin", "acne, shine", "wrinkles", "elasticity, sagging". , "Whitening, spots", "sensitive skin", "moisturizing", "price" and "comprehensive evaluation".

The "word-of-mouth ID" is the identification information of each word-of-mouth described above. The "product ID" is the above-mentioned product identification information. In addition, products are available in the fields of "pores, keratin," "acne, shine," "wrinkles," "elasticity, sagging," "whitening, age spots," "sensitive skin," "moisturizing," and "price." Numerical values representing the sentiment analysis results from "-5" to "5" described above for each function of the above are stored on average for each word of mouth. The word-of-mouth evaluation DB33-1 for each word-of-mouth sentence shown in FIG. 4 and the word-of-mouth evaluation DB33-2 for each word-of-mouth shown in FIG. 5 show examples of different data.

Next, an operation example of the processing unit 2 shown in FIG. 1 will be described with reference to FIGS. 6 to 15. FIG. 6 is a schematic diagram showing an operation example of the information processing system 1 shown in FIG. FIG. 7 is a flowchart showing an operation example of the product analysis unit 21 shown in FIG. FIG. 8 is a flowchart showing an operation example of the user analysis unit 22 shown in FIG. FIG. 9 is a flowchart showing an operation example of the word-of-mouth extraction unit 23 and the built-in unit 24 shown in FIG. FIG. 10 is a schematic diagram showing an operation example of the product analysis unit 21 shown in FIG. 11 and 12 are schematic views showing an operation example of the user analysis unit 22 shown in FIG. 1. 13 and 14 are schematic views showing an operation example of the word-of-mouth extraction unit 23 shown in FIG. FIG. 15 is a schematic diagram showing an operation example of the information processing system 1 shown in FIG.

First, with reference to FIG. 6, an outline of an operation example of the processing unit 2 shown in FIG. 1 will be described. In the processing unit 2 shown in FIG. 1, first, the product analysis unit 21 analyzes the word-of-mouth, accumulates it in the word-of-mouth evaluation DB 33, calculates the average value for each product, and stores it in the product analysis DB 31 (step ST1). Next, the user analysis unit 22 analyzes the functions that are beneficial to the user, the functions that are disadvantageous, and the functions that are not bothered (neutral functions), and accumulates them in the user attribute DB 32 (step ST2). Next, the word-of-mouth extraction unit 23 stops sending the advertisement information of the product when the negative side of the product and the function that becomes the demerit match, and evaluates the product with respect to the function that becomes the merit and the function that does not care. The word-of-mouth information extracted by the word-of-mouth extraction unit 23 is used as the print information for the delivery printed matter of envelopes, envelopes, postcards, or home deliveries. Incorporate (step ST3).

In the above operation example, in the information processing system 1, the word-of-mouth extraction unit 23 extracts word-of-mouth information included in the advertisement information related to the advertisement product for each user. At that time, the word-of-mouth extraction unit 23 includes the word-of-mouth information including a negative description about a certain function (referred to as the first function) in the word-of-mouth information of the user for the purchased product of the user in the advertisement information. Is extracted from the word-of-mouth information for the advertising product. Alternatively, in the information processing system 1, the word-of-mouth extraction unit 23 includes a word-of-mouth information related to the user's purchased product including a negative description about a certain function (referred to as the first function) of the purchased product. , Extracts information including a negative description about the first function of the advertising product from a plurality of word-of-mouth information related to the advertising product as word-of-mouth information (referred to as the first word-of-mouth information) to be included in the advertising information. According to this configuration, the negative description is likely to relate to a function that is disadvantageous to other users but not to the user. Therefore, it is unlikely that the user's willingness to purchase will be reduced. That is, according to the information processing system 1 of the present embodiment, negative word-of-mouth can be included in the advertisement information while maintaining the effect of sales promotion, and both positive word-of-mouth and negative word-of-mouth can be easily used by the user. Can be provided appropriately.

In the word-of-mouth extraction unit 23, the word-of-mouth extraction unit is word-of-mouth information including a negative description about the first function of the advertisement product from a plurality of word-of-mouth information related to the advertisement product, and the word-of-mouth information as a whole. The score (first score) according to the evaluation of the advertising product is within a predetermined range based on the score (second score) (number of stars) which is an index of the evaluation of the advertising product. It can be extracted as information.

In addition, when the word-of-mouth information related to the user's purchased product includes a positive description about the second function different from the substitute function 1 of the purchased product, the word-of-mouth extraction unit 23 uses the plurality of word-of-mouth information related to the advertising product to obtain the advertising product. Those including a positive description about the second function can be extracted as the second word-of-mouth information to be included in the advertisement information together with the first word-of-mouth information.

In addition, the word-of-mouth extraction unit 23 advertises when the word-of-mouth information related to the browsed product that the user has browsed but has not purchased contains a negative description about the first function and the third function different from the second function of the browsed product. It is possible to prevent extraction of information containing a negative description about the third function of the advertising product from a plurality of word-of-mouth information related to the product.

In addition, the incorporating unit 24 incorporates the first word-of-mouth information extracted by the word-of-mouth extraction unit 23 into the print information for the delivery printed matter of the envelope, the envelope, the postcard, or the home delivery. According to this configuration, it is possible to effectively aggregate and provide word-of-mouth in a narrow range of paper.

Next, the product analysis process of step ST1 shown in FIG. 6 will be described with reference to FIG. 7 and the like. In the process shown in FIG. 7, the product analysis unit 21 first inputs an introductory text and a word-of-mouth for the advertising product (step S11). Next, the product analysis unit 21 performs sentiment analysis on one word-of-mouth sentence written about the function for each input word-of-mouth (step S12). Next, the product analysis unit 21 performs sentiment analysis for one word-of-mouth for each input word-of-mouth (step S13). Next, in the word-of-mouth evaluation DB 33, the product analysis unit 21 stores the analysis result in step S12 in the word-of-mouth evaluation DB 33-1 for each word-of-mouth sentence shown in FIG. 4, and the analysis result in step S13 is shown in FIG. It is stored in the word-of-mouth evaluation DB 33-2 for each of the indicated word-of-mouth. Here, the word-of-mouth evaluation DB 33 stores the analysis results of word-of-mouth for the number of word-of-mouth cases.

Next, the product analysis unit 21 calculates the average value of the sentiment analysis results for each function for each product and stores it in the product evaluation DB 31 (step S14). In step S14, the product analysis unit 21 stores, for example, information 102 representing the average value of the scores indicating the evaluation results calculated based on the information 101 including the product introduction text and word-of-mouth shown in FIG. 10 in the product evaluation DB 31. ..

Next, the process of user analysis in step ST2 shown in FIG. 6 will be described with reference to FIG. 8 and the like. In the process shown in FIG. 8, the user analysis unit 22 inputs information such as age, skin type, purchase history, browsing history, etc. for each user (step S21). Next, the user analysis unit 22 determines whether or not each function of the product having a purchase history is positive (whether it is evaluated positively in the product evaluation DB 31) (step S22), and the browsing history. It is determined whether or not each function of a certain product is positive (whether or not it is evaluated positively in the product evaluation DB 31) (step S23). Then, when each function of the product having a purchase history is positive (when "yes" in step S22), or when each function of the product having a browsing history is positive (when "yes" in step S23). ), The user analysis unit 22 assigns a weight in the direction (plus direction) that is advantageous to the function (step S25), and stores the weight in the user attribute DB 32.

On the other hand, when each function of the product having a purchase history is negative (in the case of "no" in step S22), the user analysis unit 22 weights the function in the neutral direction (direction toward "0"). It is given (step S24) and stored in the user attribute DB 32. Further, when each function of the product having a browsing history is negative (in the case of "no" in step S23), the user analysis unit 22 gives a weight in a demerit direction (minus direction) to the function (step). S26), store in the user attribute DB 32.

In the process shown in FIG. 8, for example, as shown in FIG. 11, the user analysis unit 22 refers to the purchase history 111 for a certain user A, and indicates that the user A has purchased the product B and the product E. Confirm. Further, the user analysis unit 22 refers to the product evaluation DB 31, and for the product B, the function of "whitening, stain" is evaluated on the positive side, and the function of "price" is evaluated on the negative side. Make sure that. In addition, the user analysis unit 22 refers to the product evaluation DB 31, and for product E, the function of "whitening, stain" is evaluated positively, and the function of "price" and the function of "wrinkle" are negative. Make sure that the surface is evaluated. In this case, the user analysis unit 22 determines that the "whitening, spot" function is a function that is beneficial to the user A, and the "price" function and the "wrinkle" function are bothered by the user A. It is determined that the function is not (neutral function) (judgment result 113). Judging that the "price" function and the "wrinkle" function are not a concern is based on the fact that they are purchasing despite the downsides.

Further, the user analysis unit 22 refers to the browsing history 112 and confirms that the user A has browsed the product A but has not purchased it yet. In addition, the user analysis unit 22 refers to the product evaluation DB 31, and for product A, the functions of "whitening and spots" are positively evaluated, and the functions of "moisturizing" and "pores / keratin" are evaluated. Make sure that is evaluated on the negative side. In this case, the user analysis unit 22 determines that the "whitening, spot" function is a function that is beneficial to the user A, and the "moisturizing" function and the "pore / keratin" function are for the user A. Therefore, it is judged that the function is a demerit (judgment result 114). The reason why the "moisturizing" function and the "pore / keratin" function are judged to be disadvantageous functions is that it can be presumed that the downside has become a bottleneck based on the fact that they have browsed but have not purchased.

Further, in the process shown in FIG. 8, for example, as shown in FIG. 12, the user analysis unit 22 adds information 121 about the user A based on the determination result 113 and the determination result 114, and adds a function that is a merit. Points are added in the direction, points are added in the negative direction for functions that are demerits, and points are not added for functions that are not of concern (neutral functions) and are stored in the user attribute DB 32.

Next, the process of word-of-mouth extraction in step ST3 shown in FIG. 6 will be described with reference to FIG. 9 and the like. In the process shown in FIG. 9, the word-of-mouth extraction unit 23 first receives the user information (information indicating the user's attribute and the content of recognition of each function related to the user's cosmetics) stored in the user attribute DB 32. Then, a non-hierarchical cluster analysis such as the k-means method is performed, and the clusters are classified into several clusters having similar properties (step S31). Next, the word-of-mouth extraction unit 23 extracts the advantageous function, the disadvantageous function, and the neutral function for each cluster (step S32).

Next, the word-of-mouth extraction unit 23 refers to the product evaluation DB 31 and determines whether or not the negative function (negative function) of the advertising product and the demerit of each user match at least one in the same product category. Then (step S33), if they match (if "yes" in step S33), the delivery of the advertisement information is stopped (step S34).

On the other hand, if they do not match (in the case of "no" in step S33), the word-of-mouth extraction unit 23 refers to the word-of-mouth evaluation DB 33 and text-extracts the word-of-mouth written regarding the merit and neutral functions. (Step S35). Next, the word-of-mouth extraction unit 23 extracts word-of-mouth that matches the range of the overall evaluation score of the advertising product and the score of the extracted word-of-mouth (step S36). Next, the built-in unit 24 prints the word-of-mouth information extracted by the word-of-mouth extraction unit 23, for example, in the order in which the scores are more matched, for example, envelopes, envelopes, postcards, or delivery printed matter (advertising information). (Step S37), and outputs a DM (direct mail) creation and shipping instruction in which print information is printed (step S38). Note that the advertisement information is sent not only by DM but also by e-mail, SNS, or the like.

In the process shown in FIG. 9, the word-of-mouth extraction unit 23 does not care about the positive (positive expression) description of the function that is beneficial based on the information 121 of the user A, for example, as shown in FIG. The word-of-mouth 131 including a negative (negative expression) description about the function is extracted from the word-of-mouth evaluation DB 33. In this case, word-of-mouth 131 says, "I feel like the stains are getting lighter because of my mind." And "I'm glad I can feel the effect of whitening." "The price is a little easier. It includes "I wish it was a price.", And includes a positive expression about the function of "whitening, spots" and a negative expression about the function of "price".

Further, in the process shown in FIG. 9, in the word-of-mouth extraction unit 23, for example, as shown in FIG. 14, the overall evaluation score (in this case, “3.20”) for one extracted word-of-mouth 131 is an advertisement. Match the score of the evaluation index of the product (number of stars "4.20") to a certain range (for example, the score of the evaluation index "-5" to "5" to the number of stars "1" to "5" In this case, it is determined that the word-of-mouth 131 is extracted when the number of stars shown in the shaded area is within the range of "4.20 ± 0.30").

As described above, the information processing system 1 of the present embodiment creates, for example, the advertisement information 8 shown in FIG. The advertisement information 8 shown in FIG. 15 is information for advertising the beauty essence 81, and includes a score (illustration representing the number of stars) 82 of the product evaluation index and five word-of-mouth information 83 to 87. The word-of-mouth information 83 includes a word-of-mouth that "it is moist but does not like the sweet smell", and includes a positive description about "moisturizing" and a negative description about "smell". The word-of-mouth information 84 includes a word-of-mouth that "the skin has improved by one tone in a week!" And includes a positive description about "whitening and age spots". The word-of-mouth information 85 includes a word-of-mouth saying, "Even though it is a beauty essence, it has a large capacity and a good cost performance", and includes a positive description about "price". The word-of-mouth information 86 includes a word-of-mouth saying, "It is so moist for the price!", And includes a positive description about "price" and "moisture". Word-of-mouth information 87 includes a word-of-mouth saying, "It seems that the stains are getting thinner because of my mind. I wish the package was more cute.", A positive description about "whitening, stains" and about "package" Includes negative statements.

For example, when the advertisement information 8 shown in FIG. 15 is sent to a user having a neutral (don't care) user attribute for "smell" and "package", the word-of-mouth information 83 to 87 included in the advertisement information 8 is It will only include statements that are positive for the product or negative for the product but the user may not care about. In this case, it is highly probable that the word-of-mouth information 83 and the word-of-mouth information 87 including the negative description do not significantly reduce the purchase motivation of the user.

Although the embodiments of the present invention have been described above with reference to the drawings, the specific configuration is not limited to the above embodiments, and design changes and the like within a range not deviating from the gist of the present invention are also included. For example, each configuration of the information processing system 1 may be distributed, for example, to a plurality of computers via a communication line or the like. Further, the program executed by the information processing system 1 can be distributed via a computer-readable recording medium or a communication line.

1 Information processing system 2 Processing unit 21 Product analysis unit 22 User analysis unit 23 Word-of-mouth extraction unit 24 Embedded unit 3 Storage unit 31 Product evaluation DB
32 User attribute DB
33 Word-of-mouth evaluation DB
8 Advertising information 83-87 Word-of-mouth information

Claims (9)

It is an information processing system that extracts the first word-of-mouth information included in the advertising information related to the advertising products for each user.
Word-of-mouth extraction that extracts word-of-mouth information including a negative description about the first function in the user's word-of-mouth information for the purchased product of the user as the first word-of-mouth information to be included in the advertisement information from the word-of-mouth information for the advertisement product. An information processing system characterized by having a unit. The word-of-mouth extraction unit is word-of-mouth information including a negative description about the first function of the advertising product from a plurality of word-of-mouth information related to the advertising product, and responds to the evaluation of the advertising product in the entire word-of-mouth information. The information processing according to claim 1, wherein a product having a first score within a predetermined range with respect to a second score, which is an index for evaluating the advertising product, is extracted as the first word-of-mouth information. system. When the word-of-mouth information related to the purchased product of the user includes a positive description about the second function of the purchased product, the word-of-mouth extraction unit uses the second word-of-mouth information of the advertising product from the plurality of word-of-mouth information related to the advertising product. The information processing system according to claim 1 or 2, wherein a product including a positive description of a function is extracted as a second word-of-mouth information included in the advertisement information together with the first word-of-mouth information. When the word-of-mouth information related to the browsed product that the user has browsed but has not purchased contains a negative description about the third function of the browsed product, the word-of-mouth extraction unit can use the word-of-mouth information related to the advertising product. The information processing system according to any one of claims 1 to 3, wherein the third function of the advertising product is not extracted including a negative description. Any of claims 1 to 4, further comprising an embedded unit that incorporates the first word-of-mouth information extracted by the word-of-mouth extraction unit into print information for an envelope, a sealed letter, a postcard, or a delivery printed matter of a home delivery. The information processing system described in item 1. It is an information processing method that extracts word-of-mouth information included in the advertising information related to advertising products for each user.
The word-of-mouth extraction unit includes word-of-mouth information including a negative description about the first function in the user's word-of-mouth information for the user's purchased product as the first word-of-mouth information to be included in the advertisement information. An information processing method characterized by including the process of extracting from. The information processing method according to claim 6, further comprising a process of incorporating the first word-of-mouth information extracted by the word-of-mouth extraction unit into print information for a delivery printed matter of an envelope, a sealed letter, a postcard, or a home delivery. It is a program that causes a computer to execute the function of an information processing system that extracts word-of-mouth information included in the advertising information related to advertising products for each user.
The computer
Word-of-mouth extraction that extracts word-of-mouth information including a negative description about the first function in the user's word-of-mouth information for the purchased product of the user as the first word-of-mouth information to be included in the advertisement information from the word-of-mouth information for the advertisement product. A program to function as a means. The computer,
The program according to claim 8, wherein the first word-of-mouth information extracted by the word-of-mouth extraction unit functions as a means for incorporating the first word-of-mouth information into print information for a delivery printed matter of an envelope, an envelope, a postcard, or a home delivery.

Images