JP2000090094A - Information service providing device and recommendation information extracting and presenting method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To automatically extract and present information desirable for a user only by simply requesting the provision of information from the user. SOLUTION: An extraction part 131 performs the grouping of taste of respective users on retrieval conditions based on a retrieval history stored in a data base 12 and acquires the belonging coefficient of the user to each group based on the number of times of retrieval and timing. An extraction part 132 acquires the belonging coefficient of the user, who requests the position information of providing information to be printed on an information page shown by the position information request history in the data base 12 to the group, to which the user belongs, based on the belonging coefficient of that user to that group and the timing of position information request. A recommendation article extracting and presenting part 137 accepts the request of recommendation article presentation from the user and presents it to the user while determining recommendation information based on the belonging coefficient of the user to each group, to which the user belongs, and the belonging coefficients of every providing information belonging to that group.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information service providing apparatus for providing an information service to a user.
In particular, the present invention relates to an information service providing apparatus and a recommended information extracting and presenting method suitable for extracting and presenting information that matches the user's preference in response to a request from the user.

[0002]

2. Description of the Related Art In recent years, WWW (World Wide) has been developed.
Online shopping sites that sell various products on the Web) are increasing, and users (customers) who purchase products using such sites are also increasing. As the number of sites of this type increases, it tends to increase the difficulty for the user to find a desired product.

[0003] Therefore, an information service providing apparatus (information service providing server machine) having a search function of searching for a product desired by a user from among products sold on an online shopping site has been developed. This information service providing apparatus is configured to search from a vast amount of information by specifying a search condition according to a user operation from a terminal device (client machine) connected to the information service providing apparatus via a network (Internet). The product information matching the search condition is searched and presented to the user via the terminal device.

[0004] In order to enable such an information search, there is known an information service providing apparatus which makes a database of characteristics of commodities as attribute information and performs a search by designating the attribute information by a user. Here, the attribute information differs depending on the product, and if it is sake, it has attributes such as product type, type, price, and production location. In this case, in the information service providing apparatus, the user specifies a search condition consisting of an attribute and its value, for example, "the product type is Daiginjo sake, the type is dry and the price is within 5,000 yen". A process of searching for (product information) that matches the search condition from the database and presenting it to the user is performed.

[0005]

As described above, in the search for information of a product or the like by the conventional information service providing apparatus,
The user needs to specify the search condition by himself and request a search for the information he or she wants from the vast amount of information (compiled into a database). In this case, in order for the user to obtain information that suits his or her preference, search conditions must be specified accurately and finely, and there is a problem that the burden on the user is large.

Therefore, it is conceivable that the user registers his / her preference in advance in the information service providing apparatus, for example, as a user profile so that it is not necessary to specify detailed search conditions every time a search request is made. However, registering the user's own preference itself imposes a burden on the user and is not a sufficient solution. Further, it is difficult to search even information that the user himself / herself does not notice but may like by simply registering the preference.

[0007] The present invention has been made in consideration of the above circumstances, and its purpose is to simply request information provision without the user registering his / her own preference in advance, and to meet the user's preference. An object of the present invention is to provide an information service providing apparatus and a recommended information extracting and presenting method capable of automatically extracting matching information and presenting it to the user.

Still another object of the present invention is to provide an information service providing apparatus and a recommended information extracting and presenting method capable of presenting information that the user may not like but may prefer.

[0009]

According to the present invention, an information page provided by each information providing apparatus can be accessed from a user side apparatus via a network using location information indicating the location of the information page, and An information service providing apparatus applied to a system capable of information search using attribute information of provided information posted on an information page as a search condition based on an information search request with a user identifier from a side device, Based on the search condition used for the search request and the history of information search including the user identifier of the user who made the information search request, grouping of each user's preference is performed according to the search condition, and the degree of belonging of the user to each group Is obtained based on at least the number of information searches for the group at the request of the user. All user preference group data extraction means, history of a request for position information from the user device performed using the result of the information search or information page access request using the position information, and all user preference group data extraction means A preferred group recommended provision information data extraction means for obtaining affiliation degree information indicating the degree of affiliation of the provision information belonging to each group based on at least the affiliation degree information of each user with respect to the group, A request for presenting recommended information with a user identifier is received from the user device, and for each group to which the user indicated by the user identifier belongs,
Based on the affiliation degree information acquired by the all-users preference group data extraction means and the affiliation degree information of each provided information belonging to the group, based on the affiliation degree information acquired by the preference group recommended product data extraction means, Recommended information extracting and presenting means for determining and presenting the recommended information to the user who has requested the recommended information.

As described above, according to the present invention, the search condition based on the attribute information is considered to represent the user's preference, that is, what kind of provided information the search requesting user is interested in searching. The search is focused on, and user's favorite grouping is performed using the search condition. For this grouping, an information search history based on an information search request from each user is used. If the present apparatus (information service providing apparatus) has an information search service function, the search history can be obtained by the present apparatus itself, and therefore, it may be used. When the present apparatus does not have the information search service function, the search history acquired by the apparatus having the information search service function may be referred to.

When grouping is performed using search conditions, the group to which each user belongs can be determined from the search history. Here, the degree of belonging of the user to the group is considered to be greater as the number of searches performed according to the search condition of the group is larger. Based on the number of information searches for the group, for example, it can be digitized and obtained as affiliation degree information (affiliation coefficient). In addition, the degree of belonging of the user to the group can be considered to be greater as the search time is newer. Therefore, if not only the number of searches but also the search time is taken into consideration, a higher degree of belonging can be obtained. .

[0012] The user who has requested the search generally requests the position information of an information page in which certain provided information is published from the search results, or the position (if the position information is described in the search results). An information page access (reference) request based on information is issued from the user device. When such a request is issued, the user is interested in the information provided on the information page indicated by the location information, and the other users belonging to the group are also interested. It is considered that there is a high possibility that the user is interested in the provided information. Therefore, in the present invention, such provided information is regarded as belonging to the group to which the user belongs, and the degree of belonging of the provided information (listed on the information page indicated by the location information) to the group is determined as follows. It is obtained by the preferred group recommended product data extraction means based on at least the degree to which each user belongs to the group. Here, the degree of affiliation of a certain provided information with respect to a certain group is based on the fact that many users belonging to the group request a lot of position information of a page on which the provided information is provided or access to the page, and It is preferable that the larger the degree of belonging is, the larger it is. Also, the degree of belonging of the provided information to the group can be considered to be larger as the request time of the corresponding position information or the request time of access to the provided information is newer. , A higher degree of belonging can be obtained.

In order to recommend the provided information to a certain user, it is effective to recommend the provided information that many users having the same preference as the user are interested in. Therefore, in the present invention, when the user requests the presentation of the recommended information, the recommended information is determined based on the degree of belonging to each group to which the user belongs and the degree of belonging of each piece of provided information belonging to the group. This allows
The requesting user can simply present the recommended information while presenting information that suits the user's preference. Here, the higher the degree of affiliation among the groups to which the requesting user belongs, and the higher the degree of affiliation of the provision information belonging to the group to which the requesting user belongs to the group, the higher the priority as the recommended information. Good to be high.

Now, there is a correlation between each favorite group. This correlation can be considered to be stronger as the components of the corresponding search condition match. Conversely, if there is no match or a mismatch between the components, the correlation can be considered weak.

Accordingly, the present invention adds a favorite group correlation data extracting means for acquiring the degree of correlation between the groups based on the degree of matching / mismatch of the elements of the search condition corresponding to each group. By combining the degree of correlation between the favorite groups with the above-mentioned degree of belonging to the group to which the user belongs, and the degree of belonging of the provided information belonging to the group to the group, that is, each of the users to which the recommendation information presentation request source user belongs The recommended information may be determined based on the degree of affiliation to the group, the degree of affiliation of each provided information belonging to the group, and the degree of correlation of the group having a correlation with the group to the group. Features.

In such a configuration, provided information belonging to another group close to the group to which the information presentation requesting user belongs can also be a candidate for the recommended information, and the range of the candidate for the recommended information is reduced. It is also possible to present information that is not noticed by the user but may be preferred.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS [First Embodiment] First, a first embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of a system (server client system) according to the first embodiment of the present invention.

In FIG. 1, servers 10 and 20 and client 30 are connected to a network 40 such as the Internet.
It is connected to the.

The server 10 is a computer (machine) that provides information requested by the client 30, for example, information on products sold at an online shopping site.
It is. Therefore, in the following description, the server 10 is referred to as an information service providing server 10.

On the other hand, the server 20 is a computer (machine) provided for each store of the online shopping site and providing a WWW page for selling products, that is, a product page on which products are posted. In the following description, the server 20 is referred to as a general server 20 in order to clearly distinguish it from the information service providing server 10.

The client 30 is, for example, a terminal device realized using a dedicated machine or a general-purpose machine such as a personal computer. The client 30 can send a product (product information) search request and a recommended product presentation request (via the network 40) to the information service providing server 10 in accordance with a user operation.

The information service providing server 10 includes an information search service device 11, an all-user search history database 12 in which search histories of users are stored, and a recommended product extraction / presentation service device 13.

The information retrieval service device 11 receives an information retrieval request given by a user through the client 30, retrieves the requested information (here, product information), and presents the retrieval result to the request source. In the information search service device 11, the features of the products sold on the online shopping site are stored in a database as attribute information. For this reason, a search for a product in the information search service device 11 is performed by specifying attribute information of the product as a search condition.

The information retrieval service device 11 also specifies an arbitrary product from the search results for the information search request on the client 30 side, and stores the product on the W.
When a request for position information indicating the location of a WW page (product page) is sent from the client 30, the position information (for example, Uniform Resource Local) is sent.
tor, that is, URL) to the request source. By using this position information, it is possible to access (refer to) the target product page from the request source of the position information.

The information search service device 11 acquires a search history based on a search request from a user for each search request and stores the history in the all-user search history database 12. Similarly, the information search service device 11 obtains a history (position information request history) related to the position information request from the user for each position information request, and obtains the all user search history database 1
Store in 2.

FIG. 2 shows an all-user search history database 1
2 shows the details of the search history and the position information request history recorded in No. 2. Here, the search history is recorded as one record for each search request, and includes a search date and time, an identifier of the search request user (user identifier), and a search condition. The position information request history is also recorded as one record for each position information request, and includes the request date and time, the identifier of the user requesting the position information (user identifier), and the position information (URL here).

The recommended product extraction / presentation service device 13 is directly related to the present invention, and receives a recommended information presentation request (here, a recommended product presentation request) given by a user through the client 30 and matches the user's preference. The information of the recommended product is extracted and presented to the user. In order to make it possible to extract recommended products, information stored in the all-user search history database 12 is used as described later.

FIG. 3 is a block diagram mainly showing the configuration of the recommended product extraction / presentation service device 13 which is directly related to the present invention.

As shown in the figure, the recommended product extraction / presentation service device 13 includes a preference group analysis unit 130, an all-user preference group data storage unit 134, a preference group recommended product data storage unit 135, and a recommended product extraction / presentation unit 137. Have.

The preference group analysis unit 130 includes an all user preference group data extraction unit 131 and a preference group recommended product data extraction unit 132.

All user preference group data extraction unit 131
Performs a grouping of preferences, which will be described in detail below, based on search conditions based on attribute information in the search history stored in the all-user search history database 12, and indicates the degree to which each user belongs to the group. The belonging coefficient is calculated based on the search history, and the pair of the belonging coefficient and the corresponding group (group name) is recorded as favorite group data in the all-user favorite group data storage unit 134.

Here, the grouping of the preferences in the all-users preference group data extraction unit 131 will be described with reference to FIG. 4, taking the case where the product is sake as an example.

As shown in FIG. 4A, sake has attributes such as a product type, a type, a price, and a place of production. The search conditions for the information search on the sake include, for example, each attribute and its value (for a product type, for example, a product type of sake is Daiginjo sake, a type is dry and the price is within 5,000 yen). If it is a type such as Junmai Daiginjo or Ginjo sake, dry, tanrei etc.) are used. That is, FIG.
As shown in (a), all combinations of each value (attribute value) for each attribute can be search conditions.

The search condition indicated by the user in the search request can be considered to represent the user's preference as to what kind of product (information) the user is interested in searching. Therefore, in the present embodiment, favorite grouping is performed using this search condition (by attribute information).
That is, among the available search conditions, the user's preferences are grouped according to the search conditions actually specified by the user, that is, the search conditions recorded in the all-user search history database 12.

FIG. 4B shows that among the combinations of the values of the respective attributes shown in FIG.
The state of being divided into four groups of group 4 is shown. Here, a user who has made a search request according to a search condition of a favorite group is referred to as a user belonging to the group.

Next, a description will be given of the calculation of the membership coefficient indicating the degree of the membership of each user to the favorite group, that is, the extraction of the user favorite group data.

First, when the above-mentioned grouping of preferences is performed, the group to which each user belongs can be determined using the search history of the user. The user does not always request a search only under the same search condition, and may specify a different search condition from before. Therefore, a user may belong to a plurality of groups. In that case,
It can be said that the degree of belonging of each user to each group increases as the number of searches performed according to the search conditions of the group increases, and the newer the search is, the greater the degree of search becomes.

Therefore, in the present embodiment, a group belonging coefficient B (user, group) of a user as shown in the following equation (1) is introduced to determine the degree of belonging of the user (user) to the group (group). I try to digitize it.

[0040]

(Equation 1)

As described above, the belonging coefficient B (user, group) of a certain user with respect to a certain group is obtained as the sum total of the values of F (t) of the search performed for that group by the user. Here, F (t) is a monotonically decreasing function related to t, for example, 1 / t. In this case,
The affiliation coefficient (group affiliation coefficient) of a certain user for a certain group becomes larger as the number of searches for that group is larger and the search time (here, search date and time) is closer to the current time (current date and time). The degree of belonging to the group increases.

Here, a calculation example of the user's group belonging coefficient will be described. For example, from the search history in the all-user search history database 12, the number of searches performed by the user A on the group A is two, and the search date and time are:
13:00 on June 9, 1998 and 1 on May 11, 1998
It is 3:00. The current date and time is June 1998
It is assumed that it is 13:00 on March 10.

In this case, the belonging coefficient B (user A, group A) of user A to group A is Asked to do so. However, it is assumed that F (t) = 1 / t, and the unit of t is “day (da) instead of time (hour)” for simplification.
y) ".

As described above, the user does not always specify the same search condition, but may specify a different search condition from before. Therefore, a user may have belonging coefficients for a plurality of groups. FIG. 5A shows such an example. Here, an example is shown in which user A belongs to three groups A to C and has a belonging coefficient for the group.

Also, there is a possibility that a plurality of users belong to each group, and each user has a belonging coefficient for the group. FIG. 5B shows such an example.
Here, when the group is divided into group A and group B, user A and user B belong to group A, each has a belonging coefficient, and group A includes user A, user C, and user D. An example is shown in which each member has a member coefficient.

As described above, the belonging coefficient indicating the degree of belonging of the user to the group to which each user belongs is determined by the preference of all users (provided in the favorite group analysis unit 130 in the recommended product extraction / presentation service apparatus 13). The group data extracting unit 131 extracts the favorite group data as follows in the procedure according to the flowchart of FIG. 7 as follows.

First, all user preference group data extraction unit 1
31 selects an unprocessed record (next record) to be processed next from the search history recorded in the all-user search history database 12, and reads out the date and time (search date and time), the user identifier, and the search condition from the record. (Step S2).

Next, all user preference group data extraction unit 1
31 calculates F (t) based on the difference between the read date and time and the current date and time (that is, the elapsed time from the search date and time) t (step S3). Subsequently, the all-user favorite group data extraction unit 131 calculates the total sum of F (t) of the user indicated by the read user identifier with respect to (the group corresponding to) the read search condition, that is, the belonging coefficient. The value of F (t) newly calculated in step S3 is added to the intermediate calculation result (initial value 0) (step S4). This addition result (intermediate calculation result of the belonging coefficient) is held in a memory area (not shown) in pairs with the group corresponding to the search condition of the user.

All user preference group data extraction unit 131
Repeatedly executes the processing of steps S2 to S4 for all records of the search history stored in the all-user search history database 12 (step S1). Then, when all records in the search history have been executed,
The all-users preference group data extraction unit 131 uses all user preference group data as user preference group data as belonging coefficients for the groups corresponding to all search conditions of all users (recorded in the search history) that have been obtained so far. The information is recorded in the storage unit 134 (step S5).

By performing the above processing (all user favorite group data extraction processing) periodically, the latest user favorite group data is always recorded in the all user favorite group data storage unit 134.

The all user search history database 12
When the free space of a file becomes equal to or less than a certain amount, for example, the record with the oldest search date and time is deleted.
Apply rules.

In this embodiment, in addition to the above-mentioned group to which the user belongs, the concept of the group to which the product belongs is applied. Hereinafter, the group to which this product belongs will be described in detail.

The user who performed the search finds the location information (UR) of the product page on which a certain product is listed from the search results.
L) to the information service providing server 10 through the client 30 means that the user is interested in the product, and other users belonging to the group are similarly interested in the product. Can be considered to be high. That is, the product can be extracted as a recommended product of the group to which the user belongs.

In the present embodiment, such a product is regarded as belonging to the group to which the user belongs, and the product group belonging coefficient b (url, group) as shown in the following equation (2). To quantify the degree to which a product belongs to a group.

[0055]

(Equation 2)

As described above, the belonging coefficient b (url, group) for a certain group of a certain product (position information unique to)
Is a function relating to the belonging coefficient B (user, group) of all users who have requested the location information (URL) of the product page on which the product is posted and the date and time of the request for the location information (specifically, Is calculated as the total sum of the product of the time difference t between the requested date and time of the position information and the current date and time) f (t). Here, f (t) is a monotonically decreasing function related to t, for example, 1 / t. Given this, the belonging coefficient (group belonging coefficient) of a certain product for a certain group is
The greater the number of users who belong to the group, the more the position information of the page of the product is requested, and the larger the coefficient of the user's belonging to the group, the larger the value. The request time of the position information is closer to the current date and time. The greater the value, the higher the degree of the product belonging to the group.

The above-mentioned group belonging coefficient of the product is obtained by the favorite group recommended product data extraction unit 132 provided in the favorite group analysis unit 130 in the recommended product extraction / presentation service device 13 shown in FIG.

Here, an example of calculating the group belonging coefficient of a product will be described. For example, in the all-user preference group data storage unit 134, 1.03 is set as the belonging coefficient B of the user A for the group A (user A, group A), and the belonging coefficient B of the user B for the group A (user B, group A). Is assumed to be recorded as 3.11.
Also, based on the position information request history in the all-user search history database 12, the date and time when the user A requested the position information of the posting page of the product A from the search results under the search conditions corresponding to the group A is June 1998. At 13 o'clock on the 9th, the date and time when the user B requested the position information of the publication page of the product A from the search result under the search condition corresponding to the group A is 19
It is assumed that it is 13:00 on May 11, 1998. It is also assumed that the current date and time is 13:00 on June 10, 1998.

In this case, the belonging coefficient b (product A, group A) of the product A with respect to the group A is b (product A, group A) = ΣB (user, group A) f (t) = 1.03 × 1 / 1 + 3.11 x 1/30 = 1.13. However, it is assumed that f (t) = 1 / t, and for simplicity, the unit of t is “day (da) instead of time (hour)”.
y) ".

The above-mentioned requests for position information on the same product page from different users are not always made based on search results under the same search conditions, but are made based on search results under different search conditions. It is possible. Therefore, a product may have belonging coefficients for a plurality of groups. FIG. 6A shows such an example. Here, goods A are group A, group B and group C
Are shown, each of which has a belonging coefficient. Further, a plurality of products may belong to each group, and each product has a belonging coefficient for the group. FIG. 6B shows such an example. Here, when the group A and the group B exist, the product A and the product B belong to the group A, each has a belonging coefficient, and the product A, the product C, and the product D belong to the group B. Thus, an example is shown in which each has a belonging coefficient.

As described above, the belonging coefficient indicating the degree of belonging of the product to the group to which each product belongs is expressed by (the favorite group analysis unit 1 in the recommended product extraction / presentation service device 13).
The recommended group recommended product data extraction unit 132 (provided at 30) extracts the recommended group recommended product data as follows in the procedure according to the flowchart of FIG.

First, preferred group recommended product data extraction unit 1
32, an unprocessed record to be processed next (next record) is selected from the position information request history recorded in the all-user search history database 12, and date and time (request date and time), user identifier, and position information are selected from the record. Is read (step S12).

Next, all user favorite group data extraction unit 1
31 calculates f (t) based on the difference between the read date and time and the current date and time (that is, the elapsed time from the request date and time) t (step S13).

Subsequently, the all-users preference group data extraction unit 131 extracts the all-users preference group data from the all-users preference group data stored in the all-users preference group data storage unit 134.
An unprocessed group to which the user indicated by the user identifier in the record read in step S12 belongs and one belonging coefficient are read (step S15).

Next, preferred group recommended product data extraction unit 1
32 calculates the product of the value of f (t) calculated in step S13 and the belonging coefficient for the belonging group of the user read in step S15, and determines the product unique to the position information in the record read in step S12. A (partial) belonging coefficient for the group is calculated (step S1).
6). Subsequently, the preferred group recommended product data extraction unit 132
Is the sum of the belonging coefficients of the group of the products unique to the position information so far (the initial value is 0)
Then, the belonging coefficient newly calculated in step S16 is added (step S17). The result of the addition (the result of the intermediate calculation of the belonging coefficient) is stored in a memory area (not shown) of the product (position information) assigned to the group.

Preference group recommended product data extraction unit 132
The processing of the above steps S15 to S17 is performed by assigning the user belonging to the user identifier indicated by the user identifier in the record read out in step S12 among the all user favorite group data recorded in the all user favorite group data storage unit 134. It repeatedly executes the belonging coefficients of all the groups to be executed (step S14). Then, when the execution has been completed for the belonging coefficients of all the groups to which the user belongs, the favorite group recommended product data extracting unit 132 repeats the processing from step S12 on. The processing after step S12 is repeatedly executed until there is no record of the unprocessed position information request history in the all user search history database 12 (step S11).

Eventually, the all user search history database 1
When the processing of step S12 and subsequent steps is completed for all the records of the position information request history in No. 2, the favorite group recommended product data extraction unit 132 extracts all the records (recorded in the position information request history) that have been obtained so far. The affiliation coefficients of the products unique to the position information for all the favorite groups are recorded as favorite group recommended product data in the favorite group recommended product data storage unit 135 (step S18).

The above-mentioned processing (preference group recommended product data extraction processing) is performed periodically (for example, every time all user preference group data extraction processing is executed by the all user preference group data extraction unit 131), whereby preference group recommendation is performed. The product data storage unit 135 always records the latest favorite group recommended product data.

In the present embodiment, the user issues a simple recommended product presentation request with a user identifier to the client 30.
, The recommended product extraction / presentation unit 137 in the recommended product extraction / presentation service device 13 extracts and presents (information on) the product (information) that suits the user's preference. ).

As described above, in order to recommend a product to a user, it is effective to recommend a product in which many users having the same preference as the user are interested. In addition, it is more effective to recommend products that you are interested in recently than products that you were interested in in the past. The preference of a certain user is the group to which the belonging coefficient is large among the groups to which the user belongs. Also, for products that many users belonging to a group are interested in,
The belonging coefficient of the group to which the product belongs has a large value.

Therefore, the recommended product extraction / presentation unit 137
Basically, recommended products can be extracted by the following processing.

(1) The belonging coefficient for each group to which the requester of the recommended product presentation belongs belongs is detected from the all user favorite group data storage unit 134.

(2) The membership coefficient of the product belonging to each group to which the user who has requested the recommended product presentation belongs is detected from the favorite group recommended product data storage unit 135.

(3) The product of the affiliation coefficient for each group of the recommended product presentation requesting user and the affiliation coefficient (for the group) of the product belonging to the group is calculated, and the value is used as the recommended coefficient of the product. .

(4) Recommended products are selected in descending order of the value of the obtained recommended coefficient.

Next, the detailed procedure of the recommended product extraction processing by the recommended product extraction / presentation section 137 will be described with reference to the flowchart of FIG.

First, the recommended product extraction / presentation unit 137 receives a recommended product presentation request with a user identifier sent from the user via the network 40 by the client 30 (step S21).

Next, the recommended product extraction / presentation section 137 selects the group to which the user who has requested the recommended product presentation belongs and the affiliation coefficient thereof (that is, the group to which the user belongs from the all preference group data recorded in the all user preference group data storage section 134). Then, the recommended product presentation request source user belonging coefficient is read out (step S22).

Next, the recommended product extraction / presentation unit 137 selects, from the favorite group recommended product data stored in the favorite group recommended product data storage unit 135, the recommended products of the group to which the user who has requested the recommended product presentation belongs and the recommended products. The belonging coefficient (that is, the belonging coefficient of the recommended product to the group) is read (step S23).

Next, the recommended product extraction / presentation section 137 repeats the following processing for the group to which the user who has requested the recommended product presentation belongs (step S24).

That is, the recommended product extraction / presentation unit 137 first selects one (unprocessed) group to which the user who has requested the recommended product presentation belongs (step S25), and processes the recommended products of the group as follows. Is repeated (step S26).

First, the recommended product extraction / presentation unit 137 extracts one (unprocessed) recommended product of the selected group (step S27). Then, the recommended product extraction and presentation unit 137
The product of the affiliation coefficient of the recommended product presentation requesting user to the selected group and the affiliation coefficient of the extracted recommended product to the group is calculated and temporarily stored in the memory area as the recommended coefficient of the recommended product (step S2).
8).

When the steps S27 and S28 are repeated for all the recommended products in the selected group (step S26), the recommended product extraction / presentation unit 137 determines that the unprocessed group is included in the group to which the user who has requested the recommended product presentation belongs. It is checked whether or not a group exists (step S2).
4).

If there is an unprocessed group, that is, if the processing of all the groups to which the recommended product presentation requesting user belongs has not been completed, the recommended product extraction / presentation unit 137 selects the unprocessed group. One group is taken out from step S25 (step S25), and the above steps S27, S2
8 is executed.

On the other hand, when the processing of all the groups is completed, the recommended product extraction / presentation unit 137 calculates the recommended coefficient of the recommended product of all the groups to which the user who has requested the recommended product presentation belongs. , Sorts all the corresponding recommended products in descending order (in order of larger recommendation coefficient), extracts recommended products to be presented to the user, and information on the extracted recommended products, for example, products on which the recommended products are posted. The position information (URL) of the page is presented to the user through, for example, a display screen of the client 30 corresponding to the user (step S29). Here, as recommended products to be presented to the user, all recommended products may be presented in descending order of the recommended coefficient, or only recommended products whose recommended coefficients exceed the reference value may be presented in descending order of the recommended coefficient, or Even if only a certain number of recommended products are presented in descending order of the recommendation coefficient, recommended products whose recommendation coefficient exceeds the reference value may be presented in descending order of the recommendation coefficient with the fixed number as an upper limit.

FIG. 10 shows a specific example of recommended product extraction by the recommended product extraction / presentation unit 137. Here, it is assumed that the recommended product presentation request source is the user A, and the group to which the user A belongs is the groups A, B, and C.

First, the processing 101 detects the membership coefficient for the groups A, B, and C to which the user A belongs. Here, it is assumed that the belonging coefficients of the user A to the groups A, B, and C are 2.56, 3.21, and 10.28.

Next, in process 102, the belonging coefficients of the products belonging to groups A, B, and C to which user A belongs are detected. Here, it is assumed that the products 1 and 2 belong to the group A, and their belonging coefficients are 20.35 and 1.11. Also, the product 3 belongs to the group B, the belonging coefficient is 7.03, and the products 4, 5, and 5 belong to the group C.
6 belong, and their belonging coefficients are 5.25, 3.79, 0.
It is assumed to be 92.

Next, by the process 103, the belonging coefficient of the user A to each of the groups A, B, and C and the group A,
Products belonging to products B and C, that is, products of products 1 and 2 for group A, products 3 for group B and products 4 and 5 for group C are calculated, and the value is calculated as the product. The recommended coefficients are 1 to 6. Here, "commodity 1: 52.10" and "commodity 2:
2.84 "," Product 3: 22.57 "," Product 4: 5 "
3.97 "," Product 5: 38.96 "," Product 6: 9.
46 "is obtained.

Next, in the process 104, recommended products are selected in order of decreasing recommendation coefficient. Here, "Recommended Product 1: Product 4" → "Recommended Product 2: Product 1" → "Recommended Product 3: Product 5" → "Recommended Product 4: Product 3" → "Recommended Product 5: Product 6" → "Recommended" Product 6: Product 2 ”.

[Second Embodiment] Next, a second embodiment of the present invention will be described with reference to the drawings.

A feature of the second embodiment is that a configuration shown in FIG. 11 is used instead of the recommended product extraction / presentation service device 13 (having the configuration shown in FIG. 3) provided in the information service providing server 10 shown in FIG. Is used. The recommended product extraction / presentation service device 23 is different from the recommended product extraction / presentation service device 13 in that a preferred group correlation data extraction unit 233 for extracting favorite group correlation data and the favorite group correlation data are stored. A preference group correlation data storage unit 236 is added, and a recommended product extraction / presentation unit 237 for extracting recommended products in consideration of a preference group correlation is provided instead of the recommended product extraction / presentation unit 137. . The preference group correlation data extraction unit 233 includes the all user preference group data extraction unit 1
Along with 31 and the preference group recommended product data extraction unit 132 (corresponding to the preference group analysis unit 130 in FIG. 3)
The preference group analysis unit 230 is configured.

Here, the correlation between favorite groups will be described. First, there is a correlation between each favorite group. This correlation can be considered to be stronger as the components of the corresponding search condition match. vice versa,
If there is no match or no match between the components, the correlation can be considered weak.

FIG. 12A shows an example of a favorite group relating to sake. FIG. 12B shows a relationship between each of the favorite groups in the example of the favorite group in FIG. Here, for example, the group 1 and the group 2 match in three components (attributes) of the product type, the type, and the production place, and therefore, the correlation is strong. Group 1
And the group 3 are identical only in the two components of the product type and the type, and are inconsistent with each other in the place of production. Therefore, the correlation is not strong.

In the present embodiment, the degree of correlation between the respective preference groups is quantified by a correlation coefficient, and the affiliation coefficient of the user to the group to which the user belongs, and the affiliation of the products belonging to the group to the group. By combining with a coefficient, as described below, a product belonging to another group close to the group to which the recommended product presentation request source user belongs can also be set as a recommended product candidate.

Here, the correlation coefficient R (groupA, group) of group (group) B with respect to group (group) A
B) is defined as in the following equation (3).

[0097]

(Equation 3)

In the example of the favorite group in FIG. 12A, three parameters E1, E2, and E3 necessary for calculating the above R (groupA, groupB) are shown in FIG. FIG. 13B shows E2. Further, the values of R (groupA, groupB) are represented by R (group 1, group 2), R (group 1, group 3),
FIG. 13C shows four examples of R (group 2, group 1) and R (group 3, group 1).

The correlation coefficient R (group)
A, group B) is a favorite group correlation data extraction unit 2
33. Hereinafter, a detailed procedure of the favorite group correlation data extraction processing including the calculation of the correlation coefficient of the favorite group by the favorite group correlation data extraction unit 233 will be described with reference to the flowchart of FIG.

First, the favorite group correlation data extraction unit 2
33 reads all search conditions from the search history recorded in the all-user search history database 12 (step S31).

Next, the preference group correlation data extraction unit 2
33 takes out one of the read search conditions (step S33) and calculates a correlation function of all the other search conditions (corresponding to group B) with respect to the search condition (corresponding to group A) (step S34).

Preference group correlation data extraction section 233
Repeatedly executes the processing of steps S33 and S34 for all search conditions read from the all-user search history database 12 in step S31 (step S32). Then, when the execution has been completed for all the search conditions in the search history, the favorite group correlation data extraction unit 23
Reference numeral 3 stores, in the favorite group correlation data storage unit 236, correlation coefficients of all other search conditions with respect to all search conditions (recorded in the search history) obtained so far as favorite group correlation data. Yes (Step S3
5).

Now, the recommended product extraction / presentation service device 23
In the same manner as the recommended product extraction / presentation unit 137 in the previous embodiment, the recommended product extraction / presentation unit 237 extracts and presents recommended products that match the user's preference in response to a recommended product presentation request from the user. . At this time, the recommended product extraction / presentation unit 237 (unlike the recommended product extraction / presentation unit 137)
A recommended product is extracted in consideration of the correlation between the groups indicated by the favorite group correlation data recorded in the favorite group correlation data storage unit 236.

The detailed procedure of the recommended product extraction process by the recommended product extraction / presentation unit 237 will be described below with reference to FIGS.
This will be described with reference to the flowchart of FIG.

First, the recommended product extraction / presentation section 237 receives a recommended product presentation request with a user identifier sent from the user via the network 40 by the client 30 (step S41).

Next, the recommended product extraction / presentation unit 237 selects a group to which the user who has requested the recommended product presentation belongs from among all the favorite group data stored in the all user favorite group data storage unit 134 and the belonging coefficient (ie, the group to which the user belongs). The recommended product presentation request source user's belonging coefficient for the group is read (step S42).

Next, the recommended product extraction / presentation unit 237 selects, from the favorite group recommended product data stored in the favorite group recommended product data storage unit 135, the recommended products of the group to which the user requesting the recommended product belongs, and The belonging coefficient (that is, the belonging coefficient of the recommended product to the group) is read (step S43).

Next, the recommended product extraction / presentation section 237 repeats the following process for the group to which the user who has requested the recommended product presentation belongs (step S44).

That is, the recommended product extraction / presentation section 237 first selects one group to which the user who has requested the recommended product presentation belongs as the group 1 (step S45). The recommended products of the group (group 1) are as follows. The process is repeated (step S46).

First, the recommended product extraction / presentation section 237 takes out one recommended product of the selected group (step S4).
7). Then, the recommended product extraction / presentation unit 237 calculates the product of the affiliation coefficient of the recommended product presentation request source user to the selected group and the affiliation coefficient of the extracted recommended product to the group, and sets the product as the recommended coefficient of the recommended product. The data is temporarily stored in the memory area (step S48).

When the recommended product extraction / presentation unit 237 repeats the above steps S47 and S48 for all the recommended products of the selected group (step S46), the favorite group recorded in the favorite group correlation data storage unit 236 From the correlation data, all the favorite groups and correlation coefficients having a correlation with the current target group, that is, the group (group 1) selected in step S45, are read (step S49).

Next, the recommended product extraction / presentation unit 237 selects one unprocessed group from the favorite groups having a correlation read out from the favorite group correlation data storage unit 236.
One group 2 is selected (step S51), and the process is repeated as follows for the recommended products of the group (group 2) (step S52).

First, the recommended product extraction / presentation section 237 takes out one recommended product of the selected group (step S5).
3). Then, the recommended product extraction / presentation unit 237 calculates the affiliation coefficient of the recommended product presentation requesting user with respect to the group 1;
A correlation coefficient of the group 2 with respect to the group 1;
The product of the extracted recommended product with the affiliation coefficient for group 2 is calculated, and is temporarily stored in the memory area as the recommended coefficient of the recommended product (step S54).

When the recommended product extraction / presentation section 237 repeats steps S53 and S54 for all the recommended products in the group (group 2) selected in step S51 that have a correlation with group 1 (step S52), the group Another (unprocessed) group having a correlation with 1 is extracted as a group 2 (step S51), and the new group 2 is subjected to the steps S53 and S5.
Repeat 4.

Eventually, group 1 (ie, step S4
When the processes from step S51 onward are completed for all the groups having a correlation with (the group selected in step 5) (step S50), the recommended product extraction / presentation unit 237
Selects one unprocessed group to which the recommended product presentation request source user belongs as group 1 (step S4).
5) For the group (group 1) and the group having a correlation with the group (group 2), the above-described series of processing is executed.

In this manner, when the series of processing described above is completed for all the groups (group 1) to which the user who has requested the recommended product presentation belongs and the group (group 2) having a correlation with the group (group 1) Step S
44) The recommended product extraction / presentation unit 237 calculates the recommended coefficients of the recommended products of all the groups (group 1) to which the user who has requested the recommended product presentation belongs, and all the groups having a correlation with the group. According to the recommendation coefficient of the recommended product of (Group 2), all the corresponding recommended products are sorted in descending order (in order of larger recommendation coefficient) to extract recommended products to be presented to the user, and information on the extracted recommended products For example, the position information (URL) of the product page on which the recommended product is posted is presented to the user through, for example, a display screen of the client 30 corresponding to the user (step S55).

FIG. 17 shows a specific example of recommended product extraction by the recommended product extraction / presentation unit 237. Here, it is assumed that the recommended product presentation request source is the user A, the group to which the user A belongs is the group A, and the group having a correlation with the group A is the group B.

First, the process 171 detects the belonging coefficient for the group A to which the user A belongs. Here, the belonging coefficient of the user A to the group A is 2.56.
It is assumed that

Next, by the process 172, the belonging coefficient of the product belonging to the group A to which the user A belongs, the belonging coefficient of the product belonging to the group B correlated with the group A, and the group B for the group A Are detected. Here, it is assumed that the products 1 and 2 belong to the group A, the belonging coefficient is 20.35 and 1.11, the product 3 belongs to the group B, and the belonging coefficient is 7.03. I do. Also, group B for group A
Is 0.75.

Next, in the process 173, the product of the belonging coefficient of the user A to the group A and the belonging coefficient of the products 1 and 2 belonging to the group A is calculated, and the value is calculated as the recommended coefficient of the product 1 and 2. I do. Group A of user A
, The product of the correlation coefficient of group B with group A and the coefficient of belonging of product 3 with group B is calculated as the recommended coefficient of product 3. Here, "commodity 1: 52.10" and "commodity 2:
2.84 "and" Commodity 3: 13.49 "are obtained.

Next, in the process 174, recommended products are set as recommended products in descending order of the recommended coefficient. Here, the order is “recommended product 1: product 1” → “recommended product 2: product 3” → “recommended product 3: product 2”. As is apparent, the product 3 belonging to the group B having a correlation with the group A to which the user A belongs is recommended in a higher order than the product 2 belonging to the group A to which the user A belongs. It is extracted as a product. That is, the range of recommended product candidates can be expanded.

The information service providing server 10 described above
, In particular, the processing procedure of the recommended product extraction / presentation service device 13 or the recommended product extraction / presentation service device 23, more specifically, the extraction process of all user favorite group data by the preference group analysis unit 130 and the preference group recommended product A data extraction process, a recommended product extraction process by the recommended product extraction / presentation unit 137, or an all-user preference group data extraction process, a preference group recommended product data extraction process, and a preference group correlation data extraction process by the preference group analysis unit 230. The procedure of the recommended product extraction process in consideration of the preference group correlation by the product extraction / presentation unit 237 is performed by a computer readable by a program such as a CD-ROM, a floppy disk, a memory card, and the like, in which a program for executing the process procedure is recorded. Recording media And wear, can be realized by reading execute the program. This program may be loaded into a storage device of a computer through a communication medium such as a communication line.

In the above-described embodiment, a case has been described in which both the number of searches and the search time under the search conditions corresponding to the group are taken into account in calculating the user's belonging coefficient for the favorite group. It is not necessary to consider the timing of the search, although the accuracy is somewhat reduced. Similarly, it is not always necessary to consider the request timing of the position information unique to the product also regarding the belonging coefficient of the product to the favorite group.

Each client 30 is connected to the general server 2
0 is a system that accesses the product page provided by the information service providing server 10 via a specific connection device through a specific connection device, and the information service providing server 10 is placed in the connection device, the product page access using this position information (see The history may be used in place of the position information history.

Further, in the above embodiment, the case where the information service providing server 10 provides the information on the product sold on the online shopping site has been described, but the present invention provides various information other than the information on the product. The present invention can be similarly applied to the host.

[0126]

As described above in detail, according to the present invention, user preferences are grouped by using search conditions based on attribute information, and a history of information search and a history of position information request or information access request are used. Then, the degree of affiliation of each user and each piece of provided information to each group is obtained, and based on the degree of affiliation to each group to which the requesting user of the recommended information belongs and the degree of affiliation of each piece of provided information to the group, Since the information is determined, even if the user does not register his / her own preference in advance, simply requesting the provision of the information will automatically extract the information that suits the user's preference and automatically Can be presented.

According to the present invention, in addition to the degree of belonging of a user to a group to which the user belongs, and the degree of belonging of the information belonging to the group to the group, the correlation of a group having a correlation with the group to the group. By determining recommended information in consideration of the degree, provided information belonging to another group close to the group to which the information presentation requesting user belongs is also considered as a recommended information candidate, and the range of recommended information candidates is expanded. Can be.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a system according to a first embodiment of the present invention.

FIG. 2 is a diagram showing a main breakdown of a search history and a position information request history recorded in an all-user search history database 12 in FIG.

FIG. 3 is a block diagram mainly showing the configuration of a recommended product extraction / presentation service device 13 in FIG. 1;

4 is an all-user favorite group data extraction unit 1 in FIG.
The figure for demonstrating the grouping of a favorite in 31 about the case where goods are sake.

FIG. 5 is a diagram showing an example in which a user has an affiliation coefficient for a plurality of groups and an example in which a plurality of users belong to each group.

FIG. 6 is a diagram showing an example in which a product has an affiliation coefficient for a plurality of groups, and an example in which a plurality of products belongs to each group.

7 is an all-user favorite group data extraction unit 1 in FIG.
31 is a flowchart for explaining the all-user favorite group data extraction processing by 31.

8 is a preferred group recommended product data extraction unit 1 in FIG. 3;
32 is a flowchart for explaining a favorite group recommended product data extraction process by No. 32;

FIG. 9 is a flowchart for explaining recommended product extraction processing by a recommended product extraction / presentation unit 137 in FIG. 3;

FIG. 10 is a diagram showing a specific example of recommended product extraction by a recommended product extraction / presentation unit 137;

FIG. 11 is a block diagram showing a configuration of a recommended product extraction / presentation service device according to a second embodiment of the present invention.

FIG. 12 is a diagram illustrating a correlation between favorite groups.

FIG. 13 is a diagram showing a numerical example of parameters necessary for calculating a correlation coefficient indicating a degree of correlation between favorite groups, and a numerical example of a corresponding correlation coefficient.

FIG. 14 is a flowchart for explaining a favorite group correlation data extraction process by a favorite group correlation data extraction unit 233 in FIG. 11;

FIG. 15 is a diagram showing a part of a flowchart for explaining recommended product extraction processing by a recommended product extraction / presentation unit 237 in FIG. 11;

FIG. 16 is a view showing the rest of the flowchart for explaining recommended product extraction processing by the recommended product extraction / presentation unit 237 in FIG. 11;

FIG. 17 is a diagram showing a specific example of recommended product extraction by a recommended product extraction / presentation unit 237.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 10 ... Information service providing server (information service providing device) 11 ... Information search service device 12 ... All user search history database 13,23 ... Recommended product extraction and presentation service device 20 ... General server (user side device) 30 ... Client 40 ... Network 130, 230: favorite group analysis unit 131: all user favorite group data extraction unit 132: favorite group recommended product data extraction unit 134: all user favorite group data storage unit 135: favorite group recommended product data storage unit 137, 237: recommended product Extraction / presentation unit 233: Favorite group correlation data extraction unit 236: Favorite group correlation data storage unit

────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] July 26, 1999 (1999.7.2)
6)

[Procedure amendment 1]

[Document name to be amended] Statement

[Correction target item name] Claims

[Correction method] Change

[Correction contents]

[Claims]

7. An information page provided by each information providing device.
Network using location information indicating the location of the information page.
Accessible from the user device via the network, and
Information retrieval with user identifier from the user side device
Attribute of the information provided on the information page based on the request
Suitable for systems that can search for information using sex information as a search condition.
Of the recommended information extraction and presentation program used
A computer-readable recording medium that stores, in a computer, search conditions used in the information search request and the information search;
In the history of information search including the user identifier of the search requesting user
Based on the search condition, the user's favorite group
And the degree of user's belonging to each group
Affiliation degree information indicating
Based on the number of information searches for that group by request
A first step of obtaining and providing information associated with each group.
Affiliation degree information indicating the degree of affiliation of
Using the result of the search from the user-side device
Request for location information or information page using said location information
According to the access request history, at least the group
Of the information page on which the provided information belonging to
Requested access to location information or its information page
The degree to which all users belong to the group and the requirements
A second step of acquiring based on the number of requests, and presenting recommended information with a user identifier from the user-side device.
When a request is given, the user indicated by the user identifier
The first step for each group to which the
Affiliation degree information obtained by
Of each of the provided information to which the information belongs
Recommended information based on the assigned affiliation degree information
Third step of presenting the recommended information to the requesting user
Computer that stores a program for executing
Data readable recording medium. ────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] September 22, 1999 (September 9, 1999
2)

[Procedure amendment 1]

[Document name to be amended] Statement

[Correction target item name] Claims

[Correction method] Change

[Correction contents]

[Claims]

[Procedure amendment 2]

[Document name to be amended] Statement

[Correction target item name] 0009

[Correction method] Change

[Correction contents]

[0009]

According to the present invention, an information page provided by each information providing apparatus can be accessed from a user side apparatus via a network using location information indicating the location of the information page, and An information service providing apparatus applied to a system capable of information search using attribute information of provided information posted on an information page as a search condition based on an information search request with a user identifier from a side device, Based on the search condition used for the search request and the history of information search including the user identifier of the user who made the information search request, grouping of each user's preference is performed according to the search condition, and the degree of belonging of the user to each group Is obtained based on at least the number of information searches for the group at the request of the user. And the total user preferences group data extracting means, the glue of providing information belonging to each group
Affiliation degree information indicating the degree of affiliation to
The position from the user device performed using the result of the information search
Request for location information or access to information pages using location information.
At least belong to the group according to the request history
Information page on which the information provided
All users who have requested access to
The user's affiliation with the group and the number of times
A preference group recommendation providing information data extracting means for acquiring based on the number, a recommendation information presentation request with a user identifier received from the user device, and the all user preference group for each group to which the user indicated by the user identifier belongs. and affiliation degree information obtained by the data extracting means, each providing information belonging to the group, the good
And a recommended information extracting and presenting means for determining recommended information based on the affiliation degree information obtained by the group recommended provided information data extracting means and presenting the recommended information to the user who has requested the recommended information presentation. .

[Procedure amendment 3]

[Document name to be amended] Statement

[Correction target item name] 0012

[Correction method] Change

[Correction contents]

[0012] The user who has requested the search generally requests the position information of an information page in which certain provided information is published from the search results, or the position (if the position information is described in the search results). An information page access (reference) request based on information is issued from the user device. When such a request is issued, the user is interested in the information provided on the information page indicated by the location information, and the other users belonging to the group are also interested. It is considered that there is a high possibility that the user is interested in the provided information. Therefore, in the present invention, such provided information is regarded as belonging to the group to which the user belongs, and the degree of belonging of the provided information (listed on the information page indicated by the location information) to the group is determined as follows. At least the offer information belonging to the group
Information on the information page where the information is posted or its information
Group of all users who have requested access to the
Based on the degree of affiliation with the
It is obtained by only the group recommended provision information data extraction means . Here, the degree of affiliation of a certain provided information with respect to a certain group is based on the fact that many users belonging to the group request a lot of position information of a page on which the provided information is provided or access to the page, and It is preferable that the larger the degree of belonging is, the larger it is. Also, the degree of belonging of the provided information to the group can be considered to be larger as the request time of the corresponding position information or the request time of access to the provided information is newer. , A higher degree of belonging can be obtained.

Claims (8)

[Claims] An information page provided by each information providing apparatus can be accessed from a user side apparatus via a network using location information indicating the location of the information page, and an information page with a user identifier from the user side apparatus is provided. An information service providing apparatus applied to a system capable of performing an information search using attribute information of provided information posted on the information page based on an information search request as a search condition, wherein the search used in the information search request is performed. Based on the conditions and the information search history including the user identifier of the information search request source user, grouping the preferences of each user by the search condition, belonging degree information indicating the degree of belonging of the user to each group, All user preference groups obtained based on at least the number of information searches for the group at the request of the user A data extraction unit, a history of a request for the location information from the user-side device or an information page access request using the location information performed using the result of the information search, and the all-user preference group data extraction unit. A preference group recommendation provision information data extraction unit for acquiring affiliation degree information indicating the degree of affiliation of the provision information belonging to each group based on at least the affiliation degree information of each user to the group according to the data acquisition result; A request for presenting recommended information with a user identifier is received from the user-side device, and the degree-of-affiliation information obtained by the all-users-favored-group data extracting means for each group to which the user indicated by the user identifier belongs, and belonging to the group Recommended product data of the favorite group of each provided information Information service providing apparatus characterized by comprising a recommendation information extracting presentation means for presenting to a user of the recommended information presentation request source to determine the recommendation information based on the affiliation degree information obtained by the extraction means. 2. An information page provided by each information providing apparatus can be accessed from a user side apparatus via a network using position information indicating the location of the information page, and an information page provided with a user identifier from the user side apparatus. An information service providing apparatus applied to a system capable of performing an information search using attribute information of provided information posted on the information page based on an information search request as a search condition, wherein the search used in the information search request is performed. Based on the conditions and the information search history including the user identifier of the information search request source user, grouping the preferences of each user by the search condition, belonging degree information indicating the degree of belonging of the user to each group, All user preference groups obtained based on at least the number of information searches for the group at the request of the user A data extraction unit, a history of a request for the location information from the user-side device or an information page access request using the location information performed using the result of the information search, and the all-user preference group data extraction unit. A preference group recommendation provision information data extraction unit for acquiring affiliation degree information indicating the degree of affiliation of the provision information belonging to each group based on at least the affiliation degree information of each user to the group according to the data acquisition result; A preference group correlation data extraction unit that acquires correlation degree information indicating a degree of correlation between groups based on the degree of matching / mismatch of the elements of the search condition corresponding to each group; Receives a request for presentation of recommended information with a user identifier from For each group to which the indicated user belongs, the affiliation degree information obtained by the all-users preference group data extraction means and the affiliation degree of each provided information belonging to the group obtained by the preference group recommended product data extraction means Information and the correlation level information obtained by the favorite group correlation data extraction means for the group of the group having a correlation with the group, and determines recommended information based on the correlation degree information. An information service providing apparatus, comprising: a recommended information extracting and presenting means for presenting to a user. 3. The preferred group recommended product data extracting means obtains affiliation degree information of provided information belonging to each group, and includes the provided information from each user belonging to the group. 3. The method according to claim 1, wherein the number of requests for positional information of the information page or the number of access requests to the information page is used.
An information service providing device according to item 1. 4. The information search history includes information on a corresponding information search time, and the all-users preference group data extraction unit obtains information on the degree of belonging of a user to each group. 4. The information service providing apparatus according to claim 1, wherein information on a search time of an information search for the group in response to a request from the user is also used. 5. The history of the position information request or the information page access request includes information of a corresponding request time, and the favorite group recommended product data extracting means extracts the provided information belonging to each group. In order to obtain the affiliation degree information, it is necessary to use the information of the request time of the position information of the information page or the access request time to the information page on which the provided information is posted from each user belonging to the group. The information service providing device according to claim 1, wherein 6. An information page provided by each information providing apparatus can be accessed from a user side apparatus via a network using location information indicating the location of the information page, and an information page provided with a user identifier from the user side apparatus. A recommended information extraction and presentation method applied to a system capable of information search using attribute information of provided information posted on the information page as a search condition based on the information search request, the method being used for the information search request. Based on the search condition and the information search history including the user identifier of the information search request source user, grouping of the preference of each user is performed by the search condition, and affiliation degree information indicating the degree of affiliation of the user to each group is obtained. A first step of acquiring based on at least the number of times of information search for the group in response to a request from the user; According to a history of a request for the location information from the user side device or an information page access request using the location information performed using a result of the information search, and the data acquisition result in the first step, each of the groups A second step of acquiring affiliation degree information indicating the degree of affiliation of provided information belonging to at least based on the affiliation degree information of each user with respect to the group; and a request for presentation of recommended information with a user identifier from the user device. Is given, for each group to which the user indicated by the user identifier belongs, the affiliation degree information obtained in the first step and the provided information belonging to the group, The recommended information is determined based on the acquired affiliation degree information, and the user of the recommended information presentation request source is determined. Recommended information extraction presentation method characterized by comprising a third step of presenting to. 7. An information page provided by each information providing apparatus can be accessed from a user side apparatus via a network using position information indicating the location of the information page, and an information page provided with a user identifier from the user side apparatus. A recommended information extraction and presentation method applied to a system capable of information search using attribute information of provided information posted on the information page as a search condition based on the information search request, the method being used for the information search request. Based on the search condition and the information search history including the user identifier of the information search request source user, grouping of the preference of each user is performed by the search condition, and affiliation degree information indicating the degree of affiliation of the user to each group is obtained. A first step of acquiring based on at least the number of times of information search for the group in response to a request from the user; According to a history of a request for the location information from the user side device or an information page access request using the location information performed using a result of the information search, and the data acquisition result in the first step, each of the groups A second step of acquiring affiliation degree information indicating a degree of affiliation of provided information belonging to at least based on affiliation degree information of each user with respect to the group; and a correlation indicating a degree of correlation between the groups. A third step of obtaining degree information based on the degree of matching / mismatch of the elements of the search condition corresponding to each group; and a step of receiving recommended information presentation request with a user identifier from the user-side device. The location obtained in the first step for each group to which the user indicated by the user identifier belongs. The affiliation degree information, the affiliation degree information of each provided information belonging to the group, and the affiliation degree information acquired in the second step, and the group having a correlation with the group acquired in the third step for the group. A step of determining recommended information based on the correlation degree information and presenting the recommended information to the user who has requested the recommended information presentation. 8. An information page provided by each information providing apparatus can be accessed from a user side apparatus via a network using position information indicating the location of the information page, and an information page with a user identifier from the user side apparatus is provided. A computer-readable recording medium storing a recommended information extraction and presentation program applied to a system capable of performing an information search using attribute information of provided information posted on the information page as a search condition based on an information search request. Based on the search condition used in the information search request and the history of information search including the user identifier of the user who made the information search request, the computer performs grouping of each user's preference by the search condition, Attribution level information indicating the level of the user's affiliation with the group is entered at least as requested by the user. A first step of acquiring based on the number of information searches for a group; and a request for the position information from the user-side device or an information page access request using the position information performed using the result of the information search. A second step of acquiring, based on the history and the data acquisition result obtained in the first step, the degree of belonging of the provided information belonging to each group based on at least the degree of belonging of each user to the group. And, when a request for presenting recommended information with a user identifier is given from the user-side device, for each group to which the user indicated by the user identifier belongs, the belonging degree information acquired in the first step and Where each of the provided information belonging to the group is obtained in the second step. And a third step of determining the recommended information based on the attribute information and presenting the recommended information to the user who has requested the presentation of the recommended information.