JP2010113557A - Recommendation device, recommendation method and recommendation program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To achieve recommendation further promoting purchase willingness of a user. <P>SOLUTION: A commodity DB 4 of a recommendation device 1 stores articles of a commodity retrieved on an EC (Electronic commerce) site and information about contributors of the articles. A commodity retrieval part 9 refers to the commodity DB 4 about the commodity retrieved on the EC site by a user terminal 3, and acquires all the articles related to the commodity and the contributors of the articles. A recommendation control part 10 selects the contributor permitting disclosure of the article from the contributors acquired by the commodity retrieval part 9. A recommendation target person retrieval part 11 specifies the contributor corresponding to a condition designated by the user from the selected contributors. A data shaping part 12 shapes the information and the article of the contributor specified by the recommendation target person retrieval part 11 according to a shaping format 15, generates display data 16, and presents them to the user terminal 3. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a recommendation technique for automatically presenting information that meets a user's needs to a user who uses an EC site.

  In general, a site that sells products and services on the Internet is called an EC site (Electronic Commerce site). In this EC site, the user's subsequent actions after browsing a product page are divided into three patterns: (1) purchase the product, (2) search for another product, and (3) end the search. being classified.

  In order to guide users of EC sites to (1) and (2), a recommendation technique for automatically presenting information and products that meet the user's potential needs is known. In order to have more users purchase products, various recommendation technologies have been introduced on EC sites.

  As an example of the recommendation technique, a collaborative filtering method described in Non-Patent Document 1 is known. This collaborative filtering method employs a technique of analyzing the relationship between products based on the user's behavior history and displaying a product highly relevant to the product when the product is displayed.

Further, in the recommended item introduction method described in Patent Document 1, a server that performs a service that provides an item estimates client preference using client preference estimation factor information (such as an access history to an item), It introduces an item list that the client may be interested in.
JP 2002-31682 A “Classification of information recommendation methods”, http: // japan. zdnet. com / sp / feature / 08ec / story / 0, 3800085047, 20374062,00. htm

  However, when the user actually uses the EC site, when the user reaches the detail page of the product A, as described above, “The person who purchased the product A also purchases such a product” etc. Even if the product is recommended, it is not useful in determining whether or not to purchase the product A. For the user, the usability and impression (review article) of the user who actually purchased the product A is more useful information. Therefore, there are EC sites that allow users to post reviews and review articles on products.

  However, even if a word-of-mouth or review article is posted on a product, not all information is useful for the user. In particular, the weight of the impression may differ depending on whether the poster is familiar with the genre of the product or whether the product is actually purchased or used.

  For example, when importance is attached to the scarcity of the product A, it is important how much the poster is familiar with the product A or its genre. However, even if the person in charge of manufacturing who is assumed to be most familiar with product A says that product A is superior to product A of the company's product and product B of the other company's product, an objective review There are few users who can place full confidence in the content.

  Therefore, in view of such a problem, the present invention has a solution to realize a recommendation that further promotes the user's willingness to purchase.

  In order to solve the above problems, the present invention provides a condition (age, preference, etc.) of a contributor to be referred to when a user who wants to purchase a product on the Internet refers to a comment (such as a review article) of a contributor. ) Is specified, and the author's article that meets the conditions is presented.

  Specifically, the invention described in claim 1 is a recommendation device for presenting articles of products searched by a user with a terminal on a website, and a condition specified by a user from a contributor of the articles related to the searched products. And a posted article information searched by the targeted person search means and a posted article are presented on a user terminal.

  The invention according to claim 2 further includes control means for selecting a contributor who is permitted to publish a posted article to another person from a contributor of the article related to the search product, wherein the target person searching means includes: It is characterized in that a person who meets the condition is searched from the selected contributors.

  The invention described in claim 3 further includes data shaping means for shaping the relevant person's information searched by the target person search means and the posted article into a set format.

  The invention according to claim 4 is a storage means for storing an article of a product searched for on a website and information of a contributor of the article, and the storage means in conjunction with the posting of the article to the website. Update means for updating the stored information at any time, wherein the target person search means searches for a contributor who meets the condition among the searchers searched from the storage information of the storage means. Yes.

  The invention according to claim 5 is a recommendation method for presenting an article of a product searched by a user with a terminal on a Web site, wherein the target person search means is a user from an article submitter regarding the search product. It has the 1st step which searches the person who corresponds to the designated conditions, and the 2nd step which transmits the searched poster's information and the posted article to the user's terminal.

  Further, the invention according to claim 6 further includes a step in which the control means selects a contributor who has permitted others to publish the posted article from the contributors of the article related to the search product. The step is characterized by searching for a person who meets the condition from the selected contributors.

  The invention according to claim 7 is characterized in that the data shaping means further has a step of shaping the relevant person's information searched in the first step and the posted article into a set format.

  According to the eighth aspect of the present invention, the step of storing the article of the product searched on the website and the information of the contributor of the article in the storage means, and the updating means are linked to the posting of the article on the website. And updating the stored information of the storage means as needed, wherein the first step searches for a poster who meets the condition from among the posters searched from the stored information of the storage means. It is characterized by doing.

  The invention described in claim 9 is a recommendation program, and is characterized in that a computer is caused to function as each means constituting the recommendation device according to any one of claims 1 to 4.

  According to the first to ninth aspects of the present invention, since only an article of a contributor who meets a condition specified by the user is presented, a recommendation that further promotes the user's willingness to purchase is realized.

  In particular, according to the invention described in claim 2.6, the contributor of the article can arbitrarily set the publication / non-disclosure of the article.

  According to the invention of claim 3.7, the article presented to the user can be shaped into a display format suitable for the user terminal.

  Further, according to the invention of claim 4.8, an article related to a product posted by a user can be reflected in subsequent recommendations as needed.

  FIG. 1 shows a recommendation device 1 according to the present invention. The recommendation device 1 is connected to an EC site server 2 that manages the EC site via a network so as to be communicable.

  A user connects to the EC site server 2 with a user terminal 3 such as a personal computer (PC) and searches for products using the EC site provided by the EC site server 2. At this time, the recommendation device 1 extracts only the review article of the contributor who meets the condition specified by the user from the searched articles related to the product, and presents it to the user terminal 3.

  The recommendation device 1 may be a computer that can be connected to a network and can execute a program related to the recommendation, for example. Further, the recommendation device 1 is not limited to the genre of products and services handled by the EC site server 2, and any product can be applied. Further, the user terminal 3 only needs to be able to communicate with the EC site server 2 and include a display unit such as a browser, and may be a mobile terminal such as a mobile phone.

  As shown in FIG. 1, the recommendation device 1 includes a product DB 4, a product registration unit 5, a user DB 6, a user registration unit 7, a login transmission unit 8, a product search unit 9, a recommendation control unit 10, and a recommendation target person search unit. 11, a data shaping unit 12, a parameter change command unit 13, and a parameter update unit 14.

  The product registration unit 5 registers a product of an EC site recommended (recommended) to a user, a review article regarding the product, and data of a contributor thereof in the product DB 4.

  The user registration unit 7 registers, in the user DB 6, user information such as a user ID, conditions of a contributor of a review article that each user wants to browse, and the like.

  The login transmission unit 8 receives from the EC site server 2 that the user has logged in to the EC site, acquires the user-specified conditions from the user DB 6, and transmits the user-specified conditions to the recommendation target person search unit 11.

  The product search unit 9 searches the product DB 4 for a product searched by the user on the EC site, and acquires all review articles and posters of the product.

  The recommendation control unit 10 extracts contributors who are allowed to recommend to others from among the posters acquired by the product search unit 9.

  The recommendation target person search unit 11 specifies a contributor who meets the user's designated condition among the contributors extracted by the recommendation control unit 10.

  The data shaping unit 12 shapes the review article of the contributor identified by the recommendation target person searching unit 11 according to the shaping format 15 to generate display data 16 to be presented to the user, which is sent to the EC site server 2. Send back.

  The parameter change command unit 13 receives from the EC site server 2 data of review articles posted by the user on the EC site.

  The parameter update unit 14 updates the product DB 4 and the user DB 6 based on the data received by the parameter change command unit 13.

  The functions of the functional blocks 4 to 14 are realized by reading a program related to the recommendation by a control unit (CPU: Central Processor Unit, etc.) of the recommendation device 1. Further, the recommendation device 1 includes normal components of a computer, for example, an input unit such as a keyboard and a mouse (not shown), a rewritable memory (RAM) for temporarily storing processing data, and the EC site server 2. A communication device used for network connection, a storage unit such as a hard disk drive, and a display unit such as a display. Of these, each DB 4.6 is constructed on the hard disk drive. The EC site server 2 is also configured by a computer and includes the same components as described above. Hereinafter, processing executed by the functional blocks 4 to 14 will be described.

<Product DB registration process>
The merchandise master DB 21 of the EC site server 2 stores all the merchandise handled on the EC site managed by the EC site server 2 and review article data relating to each merchandise. Among these, the product to be recommended (recommended) by the user and the data of the review article of the product are transmitted from the EC site server 2 to the product registration unit 5 via the network. The product registration unit 5 registers the received product to be recommended and the data of the review article in the product DB 4. As a result, the product DB 4 stores product data of EC sites recommended by the user.

  A data example of the product DB 4 is shown in FIG. The product DB 4 includes a product list table A in which product IDs, product genres, and the like are stored, and a review table B in which review articles posted to each product and user IDs of posters are stored. Both tables A. B is associated with a product ID.

  Such registration of the product data in the product DB 4 is normally performed before the EC site is published on the network. Further, the registration of the product data may be performed by an administrator by operating the input unit. Further, the product DB 4 and the product master DB 21 may be provided in either the recommendation device 1 or the EC site server 2 as the same DB without providing them individually.

<User DB registration process>
Next, the user connects to the EC site server 2 with the user terminal 3 and performs user registration from the EC site screen provided by the EC site server 2. Here, a user ID (hereinafter referred to as EC site ID) used for logging in to the EC site, information relating to payment, and the like are registered. Such registration information is stored in the user information DB 22 provided in the EC site server 2.

  At the same time, the user registers a user ID for logging in to the recommendation device 1 (hereinafter simply referred to as a user ID). This user ID is used when the user makes various settings related to the recommendation. This user ID registration process will be described with reference to the flowchart of FIG.

  S01: The user sets a user ID from the EC site screen. This user ID may be a number or an arbitrary character string.

  S02: Next, a user profile for recommendation is set. This user profile includes fixed parameters such as the user's age (age) and gender, and variable parameters that are updated when a product is purchased. Details of the variation parameter will be described later.

  S03: Next, a default recommendation method and a recommended method are set. The recommendation method is a setting for what kind of contributor's article you want to receive when you receive a recommendation, that is, when you ask someone else to introduce a review article. Select from and set. Examples of candidates include “Purchaser”, “Same age”, “Person familiar with the genre”, “Person with similar preference”, “No recommendation required”, and the like. Here, either of them may be set as a default value.

  The recommended method is a setting for whether or not to allow a review article to be introduced to another person, that is, whether or not his / her review article is to be introduced. Here, either of them may be set as a default value.

  The settings of S01 to S03 are all performed on the EC site screen, and the data of each setting content (user ID, user profile, recommended method and recommended method) is stored in the user information DB 22 of the EC site server 2. . Further, this data is transmitted from the EC site server 2 to the user registration unit 7 via the network, and is registered in the user DB 6 by the user registration unit 7.

  Note that the user DB 6 and the user information DB 22 are synchronized, and the stored information of the same user is consistent. Moreover, you may provide both said DB as either the said recommendation apparatus 1 or the said EC site server 2, without providing both DB separately. The login to the recommendation device 1 may be authenticated by logging in to the EC site without distinguishing between the EC site ID and the user ID.

  Here, the data example of the user DB 6 is shown in FIG. The user DB 6 includes a user list table C in which a user ID, age (age), the variation parameter, and the like are stored, a purchase history for each user, and a purchase history / review article list in which review articles posted to each product are stored. A table D and a recommended / recommended method selection table E in which recommended methods and recommended methods of each user are stored. D. E is associated with the user ID.

  Here, an example of the variation parameter is shown. Here, the fluctuation parameter A = the total purchase amount of the user, the fluctuation parameter B = the genre with which the user can advise others, the fluctuation parameter C = (the number of products purchased in the set genre of the fluctuation parameter B) / (Total number Y of products purchased by each user). That is, it is indicated that a user with a high value of the variation parameter C mainly purchases products of a genre that can give advice to others.

  The variation parameters A and C are automatically updated every time a user purchases a product. The variation parameter B is input and set by the user when the user profile is set in S02.

<Examples of recommendation>
Next, an example of recommendation when a user actually searches for a product will be described based on the flowchart of FIG.

  S11-13: First, the user searches for a product on the EC site (S11), and reaches a product page as a purchase candidate (S12). Then, the EC site server 2 and the recommendation device 1 are logged in using the EC site ID and the user ID (S13). At this time, the product search in S11.12 may be performed after the login in S13.

  S14: When the user logs in to the EC site server 2, the user ID of the user is transmitted from the EC site server 2 to the login transmission unit 8. The login transmission unit 8 refers to the user DB 6 with the received user ID, and acquires the recommendation method of the user. Then, the acquired recommendation method data is transmitted to the recommendation target person search unit 11. At this time, the recommendation method desired by the user may be selected again from the EC site screen.

  S15: Next, the product ID of the purchase candidate product that the user has reached in S12 is transmitted from the EC site server 2 to the product search unit 9. The product search unit 9 refers to the product DB 4 with the received product ID, and all the review articles posted to the product from the review table B and the user IDs of the posters (hereinafter referred to as poster IDs). ) To get. And all the acquired review articles and the data of the contributor ID are transmitted to the recommendation control unit 10.

  S16: The recommendation control unit 10 extracts a contributor who permits a recommendation to others among all the contributors of the review articles received in S15. That is, the recommendation control unit 10 refers to the user DB 6 with each contributor ID of all review articles, and posts “yes” in the “introduction to others” column of the recommended / recommended method selection table E. Extract the person. The extracted contributor ID permitting the recommendation and the data of the review article are transmitted to the recommendation target person search unit 11.

  S17: The recommended target person search unit 11 identifies the contributor who corresponds to the recommendation method desired by the user received from the login transmission unit 8 in S14 among the contributors who have permitted the recommended received in S16. To do. This process will be described based on the flowchart of FIG.

  That is, the recommended person search unit 11 acquires from the recommendation control unit 10 a list of contributor IDs that are permitted to be recommended to others and a list of review articles of the contributors (S17a). This is a list (A).

  Next, the recommended person search unit 11 confirms whether or not the user's recommendation method received from the login transmission unit 8 is “person familiar with genre” (S17b). If so, the process proceeds to S17c. Otherwise, the process proceeds to S17g described later.

  Next, the recommended person search unit 11 refers to the user DB 6 with each contributor ID in the list (A), and acquires the variable parameter B of each contributor from the user list table C (S17c). .

  Then, the ID of a contributor whose genre of the obtained fluctuation parameter B matches the genre of the purchase candidate product searched by the user, that is, a poster who is familiar with the genre of the purchase candidate product, is selected and displayed in the list (B). (S17d). The number of contributors selected at this time may be one or more.

  Next, the recommended person search unit 11 extracts a contributor who matches the list (B) and the review article from the contributors in the list (A) (S17e). Then, the extracted contributor ID and the review article are targeted for recommendation, and this is set as a list (C) (S17f).

  After S17g, although there is a difference in the procedure for generating the list (B) due to the difference in the recommendation method designated by the user, the list (A) and the list (B) are similarly verified to identify the recommendation target. (S17e.17f) is performed. Therefore, only the procedure for generating the list (B) for each recommendation method will be described below.

  That is, the recommended person search unit 11 confirms whether the user's recommendation method received from the login transmission unit 8 is “same age” (S17g). If so, the process proceeds to S17h. Otherwise, the process proceeds to S17j described later.

  Next, the recommended person search unit 11 refers to the user DB 6 with each poster ID in the list (A), and acquires the age of each poster from the user list table C (S17h). Then, the ID of the poster whose acquired age matches the user's age is selected, and this is set as a list (B) (S17i).

  Further, the recommended person search unit 11 confirms whether the user's recommendation method received from the login transmission unit 8 is “purchaser” (S17j). If so, the process proceeds to S17k; otherwise, the process proceeds to S17m described later.

  Next, the recommended person search unit 11 refers to the user DB 6 with each contributor ID in the list (A), and acquires the purchase history of each contributor from the purchase history / review article list table D. (S17k). Then, in the acquired purchase history, the ID of a contributor including the purchase candidate product searched by the user is selected, and this is set as a list (B) (S17l).

  In addition, the recommended person search unit 11 confirms whether the user's recommendation method received from the login transmission unit 8 is “person with similar preferences” (S17m). If so, the process proceeds to S17n. Otherwise, the recommendation is not performed.

  Next, the recommended person search unit 11 refers to the user DB 6 with each contributor ID in the list (A), and acquires the variable parameters B and C of each contributor from the user list table C (S17n). ).

  Then, among the contributors whose fluctuation parameter B matches the genre of the candidate product for purchase searched by the user, a contributor whose value of the fluctuation parameter C is close to the value of the fluctuation parameter C of the user is selected. Let it be a list (B) (S17o).

  Here, the variation parameter C indicates the ratio of the number of product purchases in the genre of the variation parameter B to the total number of products purchased by the user. Therefore, for example, when the user purchases a product of the genre of the fluctuation parameter B at a rate of almost 100%, a contributor who purchases a product of the genre of the fluctuation parameter B at a rate of almost 100% similarly. Extracted as “person with similar preferences”.

  It should be noted that the threshold value for determining the numerical value of the variation parameter C to be close is set in advance in the program. An example of the threshold value is within 20% of the user's fluctuation parameter C within 20%. For the determination of “similar preferences”, the collaborative filtering technique of Non-Patent Document 1 may be used.

  S18: The recommendation target person search unit 11 selects the data targeted for recommendation in S17f, that is, the list (C), and transmits it to the data shaping unit 12.

  S19: The data shaping unit 12 shapes the recommendation target data in the list (C) received from the recommendation target person searching unit 11 according to the format defined in the formatting format 15, and generates display data 16 for the user. To do.

  The formatting format 15 is rule information that defines a line feed condition or a page break condition of recommendation target data to be presented to the user, and is stored in the hard disk drive device or the like.

  As an example of the display data 16, in addition to the contributor ID and the review article included in the list (C), the contributor's age (age), gender, purchase history, and the like may be added.

  S20: The data shaping unit 12 returns the display data 16 to the EC site server 2, and displays it on the EC site together with the product information (product image, etc.) held in the EC site server 2. Present to the user. A display example in this case is shown in FIG. Alternatively, the display data 16 may be transmitted to the user terminal 3 via a network and presented to the user.

  In this way, when a user who is presented with recommendation information of a product purchases the product or posts a review article regarding the product, the contents are reflected in the product DB 4 and the user DB 6. Hereinafter, processing when a review article of a product is posted by the user will be described based on the flowchart of FIG.

  S21-22: When a product review article is posted on the EC site by the user (S21), the ID of the user, the ID of the product, and the data of the posted review article are stored in the product master DB 21 of the EC site server 2. Stored in Further, the data is transmitted to the parameter change command unit 13 via the network (S22). The parameter change command unit 13 transmits the received user ID, product ID and review article data to the parameter update unit 14.

  S23 to 24: The parameter updating unit 14 refers to the product DB 4 with the product ID received in S22 (S23), and adds the content of the review article posted to the review table B and the user ID (S24).

  Further, the parameter update unit 14 refers to the user DB 6 with the user ID received in S22 (S23), and adds the content and product ID of the review article posted to the purchase history / review article list table D ( S24).

  In addition, although the case where the review article of goods was posted by the user was demonstrated here, also when the user purchases goods or when the user's age (age) is updated, both said DB4. The six related parameters are updated as well. In this way, by receiving the parameter change trigger from the EC site server 2 as needed, both the DBs 4.6 are updated to the latest state.

  As described above, according to the recommendation device 1, since only the review article of the contributor that meets the condition specified by the user is presented, the recommendation that further promotes the user's willingness to purchase can be realized. .

  Moreover, since only the review article of the contributor who meets the desired condition is recommended for the user, it is possible to easily determine whether or not to purchase the product. Moreover, since the conditions of the contributor who desires recommendation can be changed at any time, the user can receive the recommendation suitable for various situations.

  Furthermore, by browsing the posting history and purchase history of contributors with similar preferences, the user can easily find a product that satisfies his / her potential needs.

  In addition, this invention is not limited to the said embodiment, For example, it can provide also as a recommendation program which makes a computer function as each functional block 4-14 of the said recommendation apparatus 1. FIG. This program may cause the computer to realize all the functions of the functional blocks 4 to 14, or may realize a part of the functions.

  This program is provided to the computer by downloading from a website or the like. The program is stored in a recording medium 10 such as a CD-ROM, a DVD-ROM, a CD-R, a CD-RW, a DVD-R, a DVD-RW, an MO, an HDD, or a Blu-ray Disk (registered trademark). May be provided to a computer.

The block diagram of the recommendation apparatus which concerns on embodiment of this invention. Data example of the product DB4. The user registration process flow figure to the recommendation apparatus 1. FIG. Data example of the same user DB6. The processing flow figure of the recommendation Example. The processing flow figure of the recommendation subject search part 11. FIG. Display example of the recommendation information. The update process flow figure of the goods DB4 and user DB6.

Explanation of symbols

1 ... Recommendation device 2 ... EC site server 21 ... Product master DB
22 ... User information DB
3 ... User terminal 4 ... Product DB (storage means)
5. Product registration unit 6. User DB
DESCRIPTION OF SYMBOLS 7 ... User registration part 8 ... Login transmission part 9 ... Merchandise search part 10 ... Recommendation control part 11 ... Recommendation target person search part 12 ... Data shaping part 13 ... Parameter change command part 14 ... Parameter update part (update means)
15 ... Formatting format 16 ... Display data A ... Product list table B ... Review table C ... User list table D ... Purchase history / review article list table E ... Recommended / recommended method selection table

Claims (9)

A recommendation device that presents articles about products searched by a user on a website on a website,
From a contributor of an article related to the search product, comprising a target person search means for searching for a person who meets a user-specified condition,
The recommendation apparatus characterized by presenting the information of a contributor searched by the target person search means and the posted article on a user terminal. From the contributor of the article related to the search product, further comprising a control means for selecting a contributor who is permitted to publish the posted article to others,
The recommendation apparatus according to claim 1, wherein the target person search unit searches for a person who meets the condition from the selected contributors. The recommendation device according to claim 1, further comprising data shaping means for shaping the information of the corresponding person searched by the target person search means and the posted article into a set format. Storage means for storing articles of products to be searched on a website and information of contributors of the articles;
Update means for updating the storage information of the storage means at any time in conjunction with the posting of an article on a website,
The said object person search means searches for the contributor who meets the said condition among the contributors searched from the storage information of the said memory | storage means. The Claim 1 characterized by the above-mentioned. Recommendation device. A recommendation method for presenting articles of products searched by a user with a terminal on a website,
A first step of searching for a person who meets a user-specified condition from a contributor of an article about the search product;
And a second step of transmitting the searched poster information and posted articles to the user's terminal. The control means further comprises a step of selecting a contributor who has permitted others to publish a posted article from the contributor of the article related to the search product,
The recommendation method according to claim 5, wherein the first step searches for a person who meets the condition from the selected contributors. The recommendation method according to any one of claims 5 and 6, further comprising a step of shaping the information of the corresponding person searched in the first step and the posted article into a set format. . Storing the article of the product searched on the website and the information of the poster of the article in the storage means;
The update means further includes the step of updating the stored information of the storage means as needed in conjunction with the posting of an article to a website;
The recommendation according to any one of claims 5 to 7, wherein the first step searches for a contributor who meets the condition among contributors searched from the storage information of the storage unit. Foundation method.   A recommendation program for causing a computer to function as each means constituting the recommendation device according to claim 1.

Images