JP2004341584A - System, method, program, and recording medium for introducing information - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device, a method, a program and a recording medium for introducing information, which provide highly accurate introduction information even if a history of access to Web sites, stores, users and the like is not enough. <P>SOLUTION: A history of accesses by a number of users is retained in an integral access history retaining means. A second group of items highly correlated with a predetermined item received from a history analysis introduction device is extracted from the integral access history retaining means. A third group of items highly correlated with an item included in the first group of items which is highly correlated with the item accessed by a user is extracted from the integral access history retaining means. From the items included in the second and third groups of items, a fourth group of items is extracted based on a predetermined standard and sent as part of recommended items to be introduced to the predetermined user. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to an information introduction system, an information introduction method, a program, and a recording medium for introducing recommended items based on an accelerator history in which many users have accessed each of a large number of items.
[0002]
[Prior art]
Websites and stores that handle a lot of products and information categorize those products and information into various genres, and prepare specialty products and special offers in each of the above genres. , Products, and information. Also, among the classified products, present products that are suitable for each user as recommended products, and when issuing a discount coupon, present products and information that the user may be more interested in. ing.
[0003]
When it is desired to know where the interest of a predetermined user is, generally, the predetermined user is asked to specify a genre of interest, but in this method, detailed interest can be grasped. Have difficulty. Therefore, there is a problem that when a predetermined user is designated a genre of interest, individual products cannot be introduced within a specific genre.
[0004]
Conventionally, as a method of knowing the user's interest, a method of knowing the user's interest by using an access history to each product has been known. In this conventional method, since the analysis is performed for each product, it is possible to introduce products recommended for individual products.
[0005]
In the case of performing an analysis on a product basis, a collaborative filtering technique is often used as the analysis method.
[0006]
The above-mentioned collaborative filtering method compares the access history of a predetermined user with the access history of another user, detects a user having a similar access history, and detects the user having the similar access history. Indicates the relationship between the product accessed by the detected user and the product accessed by the predetermined user, and indicates the relationship between the product accessed by the detected user and the predetermined user. Method.
[0007]
In this collaborative filtering method, if there is a history of a predetermined user accessing several products from among the entire products, products based on those access histories can be introduced (for example, Patent Documents 1 and 2). , Non-Patent Documents 1 to 7).
[0008]
[Patent Document 1]
Japanese Patent Application Laid-Open No. 2002-082641 [Patent Document 2]
Japanese Patent Application Laid-Open No. 2002-298007 [Non-Patent Document 1]
Takeshi Motohashi 2 "Advertising / Marketing Technology in Portal Services" NTT Technical Journal Vol. 14, No. 1, pp85-87, 2002
[Non-patent document 2]
Takeshi Motohashi 1 person "Awareness net supporting e-business development based on access history" NTT Technical Journal Vol. 12, no. 11, pp84-85, 2000
[Non-Patent Document 3]
Takeshi Motohashi 1 "Awareness net that analyzes and recommends user's preference only from access history" IPSJ, 62nd Annual Conference Special Track (1) Groupware and Network Services, National Conference 253-256 March 2001 [Non-Patent Document 4]
Takanori Ukai et al. "Directory management using collaborative filtering of bookmarks" Information Processing Society of Japan research groupware and network services, Vol. 44, no. 9, pp. 45-50, May 2002 [Non-Patent Document 5]
Keiichiro Hoashi et al., "Search method expansion method applying evaluation value prediction algorithm in collaborative filtering, Transactions of Information Processing Society of Japan, Vol. 43, No. 4, April 2002 [Non-Patent Document 6]
D. Goldberg Outside 3 Using Collaborative Filtering to Weave an Information Tapestry Communications of the ACM, Vol. 35, no. 12, pp. 51-60, December 1992
[Non-Patent Document 7]
Social Information Filtering Algorithms for Automating “Word of Mouth” Proceedings of the ACM Conference on Human Resource Company 210-217, May 1995
[0009]
[Problems to be solved by the invention]
However, the above-described conventional collaborative filtering method has a problem that if there is little original information, the accuracy of introducing recommended items (information such as product information, event information, and document information) is low.
[0010]
An object of the present invention is to provide an introduction device, an introduction method, an introduction program, and a recording medium that can provide highly-accurate introduction information even when an access history in a website, a store, a user, or the like is small. Things.
[0011]
[Means for Solving the Problems]
According to the present invention, a history of access by a large number of users is held in an integrated access history holding unit, and a second item group that is highly correlated with a predetermined item received from the history analysis and introduction device is integrated with an integrated access history. A third item group that is extracted from the holding unit and is an item that is highly correlated with an item included in the first item group that is highly correlated with the item accessed by the user is extracted from the integrated access history holding unit. , Extracting a fourth item group based on a predetermined criterion from among the items included in the second and third item groups, and transmitting the extracted item group as a part of recommended items to be introduced to a predetermined user It is.
[0012]
Embodiments and Examples of the Invention
FIG. 1 is a diagram showing an information introduction system 100 according to one embodiment of the present invention.
[0013]
The information introduction system 100 includes a history analysis introduction device 10 and a history integrated analysis introduction device 20.
[0014]
The history analysis and introduction device 10 includes a local access history holding unit 11 and a history introduction unit 12.
[0015]
The history integrated analysis introduction device 20 includes an integrated access history holding unit 21 and an integrated history introduction unit 22.
[0016]
The local access history holding unit 11 is a unit that holds, for each user, a history of a predetermined user U accessing a predetermined item.
[0017]
The history introduction unit 12 has a first item group extraction unit and a transmission unit.
[0018]
The first item group extracting means is a means for extracting a first item group I1 which is an item having a high correlation with the predetermined item A accessed by the predetermined user U.
[0019]
The transmitting means is a means for transmitting the predetermined item A and the first item group I1.
[0020]
The integrated access history holding unit 21 is a unit that holds histories of access by many users other than the predetermined user U.
[0021]
The integrated history introduction unit 22 is a unit that includes a second item group extraction unit, a third item group extraction unit, a fourth item group extraction unit, and a fourth item group transmission unit.
[0022]
The second item group extraction unit is a unit that extracts, from the integrated access history holding unit 21, a second item group I2 that is an item having a high correlation with the predetermined item A received from the history analysis introduction device 10.
[0023]
The third item group extracting unit is a unit that extracts, from the integrated access history holding unit 21, a third item group I3 that is highly correlated with the items included in the first item group I1.
[0024]
The fourth item group extracting means selects the fourth item group based on a predetermined criterion from among the items included in the second item group I2 and the items included in the third item group I3. This is a means for extracting groups.
[0025]
The fourth item group transmitting unit is a unit that transmits the fourth item group I4.
[0026]
In addition, the history analysis introduction device 10 receives the fourth item group I4 transmitted by the fourth item group transmission unit, and among the items obtained by adding the first item group I1 and the fourth item group I4. , A device for introducing a desired number of items to the user U as a recommended item group.
[0027]
In addition, the history integrated analysis and introduction device 20 periodically receives access history transfer from devices other than the history analysis and introduction device 10.
[0028]
The predetermined criterion is a criterion for extracting a predetermined number from items having a high number of appearances among items included in the second item group I2 and items included in the third item group I3. is there.
[0029]
The predetermined criterion is a criterion for preferentially extracting an item that appears in common between an item included in the second item group I2 and an item included in the third item group I3. There may be.
[0030]
Next, the operation of the information introduction system 100 will be described more specifically.
[0031]
That is, an operation when the information introduction system 100 is applied to an information providing service on the Internet will be described.
[0032]
That is, when the user U refers to the history analysis / introduction device 10 provided on a predetermined Web site using a personal computer, the user U sends the access history by the user U to the history integration analysis / introduction device 20, and The operation when receiving a recommended item introduction to U will be described.
[0033]
FIG. 2 is a flowchart showing the operation of the above embodiment.
[0034]
First, when a predetermined user U accesses a predetermined item A, information of the accessed item A is transmitted to the history analysis and inquiry service system 10 (S1). In this case, the item A is various information such as product information, event information, document information, personal network information, stock information, and the like. Then, a recommended item group (first item group I1) (a plurality of items) having a high correlation with the item A purchased by the user U is extracted from the local access history holding unit 11 (S11).
[0035]
Here, it is assumed that the history analysis introduction device 10 wants to introduce five items to the user U as recommended items. If the number of items included in the extracted first item group I1 is 5 or more (S12), a predetermined criterion (for example, a criterion for introducing an item having a large number of appearances) is used. The five items are introduced to the user U (S13).
[0036]
However, if the number of items included in the extracted first item group I1 is less than 5 (S12), the item A and the first item group I1 having a high correlation with the item A are determined as follows: The history integration analysis introduction device 20 is notified and requested to introduce a recommended item (S14).
[0037]
The history integrated analysis introduction apparatus 20 first extracts the second item group I2 having a high correlation with the item A from the integrated access history holding unit 21 (S21). Further, a third item group I3 having a high correlation with each item included in the first item group I1 is extracted (S22). Then, among the items included in the second item group I2 and the items included in the third item group I3, an item having a large number of appearances is extracted as a fourth item group I4 ( S23).
[0038]
The extracted fourth item group I4 is transmitted to the history analysis and introduction device 10 (S24).
[0039]
Note that the history analysis and introduction device 10 that has received the fourth item group I4 has added the first item group I1 (the first recommended item) having a high correlation with the item A and the fourth item group I4. Among the items, five items are set as a recommended item group (S30).
[0040]
In this case, items having a large number of appearances are extracted from the fourth item group I4 by the number obtained by subtracting the number of items included in the first item group I1 from “5”. An item to which the items included in the first item group I1 are added is referred to as a recommended item group.
[0041]
In addition, among a plurality of items included in the first item group I1 and the fourth item group I4, five items are selected in descending order of the number of appearances, and the selected items are recommended. It may be an item group.
[0042]
Then, the history analysis introduction device 10 transmits the recommended item group to the user U (S31).
[0043]
In the step S23, the fourth embodiment is used as a reference for extracting the fourth item group I4 from the items included in the second item group I2 and the items included in the third item group I3. Extracts "the item having a large number of appearances among the items included in the second item group I2 and the items included in the third item group I3". Instead, a criterion for extracting “items common to the items included in the second item group I2 and the items included in the third item group I3” may be adopted. It may be.
[0044]
That is, the item list of the second item group I2 having a high correlation with the item A is:
・ Item B
・ Item C
・ Item D
・ Item F
・ Item K
And
[0045]
Also, a list of items of the third item group I3 having a high correlation with the items included in the first item group I1 is as follows:
・ Item A
・ Item D
・ Item E
・ Item K
And
[0046]
The items included in both the item lists in common are two items, item D and item K, and the items D and K are extracted as a fourth item group I4.
[0047]
The information introduction system 100 is an apparatus for introducing information using similarity calculation between items. The “calculation of similarity between items” is called social filtering (hereinafter referred to as “SF”), is widely known, and is also used in Non-Patent Documents 1 to 7 described above.
[0048]
Next, the operation of introducing information about SF (social filtering) in the information introduction system 100 will be described.
[0049]
FIG. 3 is a diagram for explaining the operation of introducing SF (social filtering) information in the information introduction system 100.
[0050]
The information that the user U0 has referred to most recently (eg, one to two days ago, one to two weeks ago, etc.) is referred to as the most recent behavior information group G1, and information having a high degree of similarity to the most recent behavior information group G1 is used. The group of users is referred to as a related user group G2. Note that the related user group G2 does not include the user U0. Information that is included in the related user group G2 and has been referred to by the users and is not included in the information group G1 is referred to as an introduction behavior information group G3.
[0051]
Since the introduction behavior information group G3 has a high similarity to the information referred to by the user U0, it is estimated that the user U0 is interested in the introduction behavior information group G3, and the introduction behavior information group G3 is provided to the user U0. To introduce.
[0052]
Here, a value indicating the degree of similarity is assigned to each related user included in the related user group G2, and the value indicating the degree of similarity is used for the introduction behavior information group G3. A value indicating the estimated degree of interest of each piece of information included is given. For example, they are introduced in descending order of the estimated interest degree.
[0053]
In the above embodiment, the number of recommended item groups may be set to a plurality other than five.
[0054]
Further, in the above embodiment, the user U accesses the item A to the website or the like, but the above embodiment is applied when the user U purchases the item A. be able to.
[0055]
Further, in the above embodiment, the history analysis and introduction device 10 is installed on a Web site. However, when the history analysis and introduction device 10 is installed in a store, the above embodiment can be applied. .
[0056]
In the above embodiment, since the history integrated analysis and introduction device 20 periodically acquires the access history from the devices other than the history analysis and introduction device 10, the Web site and the store with a small access history extract highly accurate introduction information. be able to. Therefore, even at a Web site or a store with a small access history, information matching the user's interest can be provided.
[0057]
Further, according to the above embodiment, a large amount of access history acquired in the store service can be provided in response to inquiries from individual websites and stores, so that new business development is possible.
[0058]
The above embodiment can be understood as a program invention.
[0059]
That is, in the above embodiment, the receiving procedure of receiving the predetermined item accessed by the predetermined user and the first item group that is highly correlated with the predetermined item from the history analysis and introduction device, An integrated access history holding procedure for storing the history of accesses by users other than the predetermined user in the integrated access history holding means, and a second item which is highly correlated with the predetermined item received from the history analysis and introduction device. A second item group extraction procedure for extracting the item group from the integrated access history holding means, and a third item group that is highly correlated with the items included in the first item group. A third item group extraction procedure to be extracted from the integrated access history holding unit; an item included in the second item group; A fourth item group extraction procedure for extracting a fourth item group based on a predetermined criterion among the items included in the group, and introducing the fourth item group to the predetermined user. 13 is an example of a program for causing a computer to execute a fourth item group transmission procedure to be transmitted as a part of recommended items to be transmitted.
[0060]
Further, the program may be recorded on a recording medium such as a CD, DVD, HD, or semiconductor memory.
[0061]
【The invention's effect】
ADVANTAGE OF THE INVENTION According to this invention, there exists an effect that the information which matched the user's interest can be provided appropriately also in Web site or store with a small access history.
[Brief description of the drawings]
FIG. 1 is a diagram showing an information introduction system 100 according to an embodiment of the present invention.
FIG. 2 is a flowchart showing the operation of the embodiment.
FIG. 3 is a diagram for explaining the operation of introducing information about SF (social filtering) in the information introduction system 100.
[Explanation of symbols]
100… Information introduction system,
10 ... history analysis introduction device,
11 local access history holding means,
12 ... History introduction means,
20… History integrated analysis introduction device,
21 ... Integrated access history holding means,
22 ... Integration history introduction means.

Claims (11)

An information introduction system including a history analysis introduction device and a history integrated analysis introduction device,
The history analysis introduction device is
Local access history holding means for holding, for each user, a history of a predetermined user accessing a predetermined item;
First item group extracting means for extracting a first item group which is an item having a high correlation with the predetermined item accessed by the predetermined user;
Transmitting means for transmitting the predetermined item and the first item group;
Has,
The history integration analysis introduction device is
Integrated access history holding means for holding a history of accesses by users other than the predetermined user;
Second item group extraction means for extracting, from the integrated access history holding means, a second item group that is an item having a high correlation with the predetermined item received from the history analysis introduction device;
Third item group extracting means for extracting, from the integrated access history holding means, a third item group that is highly correlated with the items included in the first item group;
A fourth item group that extracts a fourth item group based on a predetermined criterion from among the items included in the second item group and the items included in the third item group Extraction means;
Fourth item group transmission means for transmitting the fourth item group;
The history analysis and introduction device receives the fourth item group transmitted by the fourth item group transmission unit, and stores an item obtained by adding the first item group and the fourth item group. An information introduction system, wherein a desired number of items are introduced to the user as a recommended item group. In claim 1,
An information introduction system, wherein the history integrated analysis introduction device receives an access history periodically from a device other than the history analysis introduction device. In claim 1,
The predetermined criterion is a criterion for extracting a predetermined number from items having a large number of appearances among items included in the second item group and items included in the third item group. An information introduction system characterized by the following. In claim 1,
The predetermined criterion is a criterion for preferentially extracting an item commonly appearing between an item included in the second item group and an item included in the third item group. Information introduction system characterized by the following. Receiving means for receiving, from the history analysis and introduction device, a predetermined item accessed by a predetermined user and a first item group that is highly correlated with the predetermined item;
Integrated access history holding means for holding a history of accesses by users other than the predetermined user;
Second item group extraction means for extracting, from the integrated access history holding means, a second item group that is an item having a high correlation with the predetermined item received from the history analysis introduction device;
Third item group extracting means for extracting, from the integrated access history holding means, a third item group that is highly correlated with the items included in the first item group;
A fourth item group that extracts a fourth item group based on a predetermined criterion from among the items included in the second item group and the items included in the third item group Extraction means;
A fourth item group transmitting means for transmitting the fourth item group as a part of a recommended item to be introduced to the predetermined user;
A history integrated analysis introduction device characterized by having: In claim 5,
The history integrated analysis and introduction device, wherein the access history is periodically transferred from a device other than the history analysis and introduction device. In claim 5,
The predetermined criterion is a criterion for extracting a predetermined number from items having a large number of appearances among items included in the second item group and items included in the third item group. A history integration analysis introduction device characterized by the following. In claim 5,
The predetermined criterion is a criterion for preferentially extracting an item commonly appearing between an item included in the second item group and an item included in the third item group. A history integration analysis introduction device characterized by the following. Receiving a predetermined item accessed by a predetermined user and a first group of items that are highly correlated with the predetermined item from the history analysis and introduction device;
An integrated access history holding step of holding a history of accesses made by users other than the predetermined user in an integrated access history holding means;
Extracting a second item group, which is an item having a high correlation with the predetermined item received from the history analysis and introduction device, from the integrated access history holding means;
Extracting a third item group, which is an item having a high correlation with the items included in the first item group, from the integrated access history holding means;
A fourth item group that extracts a fourth item group based on a predetermined criterion from among the items included in the second item group and the items included in the third item group An extraction step;
Transmitting a fourth item group as a part of recommended items to be introduced to the predetermined user;
An information introduction method characterized by having: A receiving procedure of receiving, from the history analysis and introduction device, a predetermined item accessed by a predetermined user and a first item group that is highly correlated with the predetermined item;
An integrated access history holding procedure for storing the history of accesses by users other than the predetermined user in the integrated access history holding means;
A second item group extraction procedure for extracting, from the integrated access history holding means, a second item group that is highly correlated with the predetermined item received from the history analysis introduction device;
A third item group extraction procedure for extracting, from the integrated access history holding means, a third item group that is highly correlated with the items included in the first item group;
A fourth item group that extracts a fourth item group based on a predetermined criterion from among the items included in the second item group and the items included in the third item group Extraction procedure;
Transmitting a fourth item group as a part of a recommended item to be introduced to the predetermined user;
A program that causes a computer to execute. A receiving procedure of receiving, from the history analysis and introduction device, a predetermined item accessed by a predetermined user and a first item group that is highly correlated with the predetermined item;
An integrated access history holding procedure for storing the history of accesses by users other than the predetermined user in the integrated access history holding means;
A second item group extraction procedure for extracting, from the integrated access history holding means, a second item group that is highly correlated with the predetermined item received from the history analysis introduction device;
A third item group extraction procedure for extracting, from the integrated access history holding means, a third item group that is highly correlated with the items included in the first item group;
A fourth item group that extracts a fourth item group based on a predetermined criterion from among the items included in the second item group and the items included in the third item group Extraction procedure;
Transmitting a fourth item group as a part of a recommended item to be introduced to the predetermined user;
Computer-readable recording medium on which a program for causing a computer to execute the program is recorded.

Images