Service providing apparatus and method for cross-recommendation between product sales sites based on online, service providing system therefore, and non-transitory computer readable medium having computer program recorded thereon

1 . A service providing apparatus for cross-recommendation between online product sales sites, the service providing apparatus comprising:

one or more processors; and

memory storing instructions thereon that, when executed by the one or more processors, cause the one or more processors to:

collect behavior log information of a plurality of users for each of a plurality of different product sales sites, each product sales site being a different web site;

group items of the behavior log information into log groups such that each log group comprises behavior log information corresponding to a common user and a common category of a product sales site;

calculate a category point for each log group by applying event types included in the behavior log information of the log group to a first algorithm that assigns a respective weight to each event type;

create, for each log group, category preference information comprising category identification information identifying the user, the product sales site, and the category corresponding to the log group, and the category point;

calculate similarity between categories of different product sales sites by applying a second algorithm to the category preference information created for a plurality of users;

create matching information in which categories of different product sales sites are matched to each other based on the similarity; and

generate recommendation result information comprising a category of a first product sales site matched, based on the matching information, to a category of a second product sales site preferred by a specific user, and provide the recommendation result information to a user terminal of the specific user connecting to the first product sales site.

2 . The service providing apparatus of claim 1 , wherein the first algorithm is collaborative filtering.

3 . The service providing apparatus of claim 1 , wherein the behavior log information comprises information about at least one of product search, adding, putting in a cart, and buying by the user.

4 . The service providing apparatus of claim 1 , wherein the second algorithm is a cosine distance, and the instructions further cause the one or more processors to calculate a distance according to the cosine distance between a plurality of different categories based on the category preference information, use the calculated distance as the similarity, and create the matching information by matching, for each of the plurality of different categories, a category of a product sales site to one or more categories of another product sales site having the highest similarity.

5 . The service providing apparatus of claim 1 , wherein the instructions further cause the one or more processors to:

select a category of the second product sales site preferred by the specific user as a preference category based on one or more items of the behavior log information corresponding to the second product sales site and the specific user using a recommendation algorithm; and

identify a category of the first product sales site matched to the preference category as a recommendation category based on the matching information, and generate the recommendation result information for recommending a product pertaining to the recommendation category.

6 . The service providing apparatus of claim 5 , wherein the instructions further cause the one or more processors to:

receive connection information of the specific user connecting to the first product sales site from an affiliated store server corresponding to the first product sales site;

select one or more interest products preferred by the specific user from products sold on the second product sales site based on the behavior log information corresponding to the specific user and the second product sales site using the recommendation algorithm;

determine one or more recommendation categories by selecting a category to which the one or more interest products pertain as the preference category from categories of the specific of the second product sales site and identifying a category of the first product sales site matched to the preference category as a recommendation category based on the matching information; and

transmit the recommendation result information comprising the one or more recommendation categories to the affiliated store server for the affiliated store server to select and recommend a recommendation product to the specific user based on the one or more recommendation categories, or transmit product recommendation information comprising one or more recommendation products to the affiliated store server or the user terminal of the specific user.

7 . The service providing apparatus of claim 6 , further comprising a storage unit storing product information received from each of the plurality of product sales sites, wherein the instructions further cause the one or more processors to:

check whether a visit history of the specific user exists for each of the first product sales site and the second product sales site based on the behavior log information corresponding to the specific user upon receiving the connection information;

responsive to a visit history of the specific user existing on the first product sales site and the second product sales site, select a recommendation product from product information corresponding to the first product sales site stored in the storage unit based on the behavior log information corresponding to the first product sales site and the specific user, select a recommendation product for each of one or more recommendation categories according to the recommendation result information from the product information corresponding to the first product sales site, and create product recommendation information comprising the selected recommendation products; and

responsive to a visit history of the specific user existing on the second product sales site and an absence of a visit history on the first product sales site, create product recommendation information comprising a recommendation product selected according to the recommendation result information from the product information corresponding to the first product sales site.

8 . The service providing apparatus of claim 1 , wherein the instructions further cause the one or more processors to create the matching information by, for each category of a product sales site, excluding categories of the same product sales site and matching the category to one or more categories of a different product sales site having the highest similarity.

9 . A computer-implemented method for cross-recommendation between online product sales sites, the method comprising:

collecting behavior log information of a plurality of users for each of a plurality of different product sales sites, each product sales site being a different web site;

grouping items of the behavior log information into log groups such that each log group comprises behavior log information corresponding to a common user and a common category of a product sales site;

calculating a category point for each log group by applying event types included in the behavior log information of the log group to a first algorithm that assigns a respective weight to each event type;

creating, for each log group, category preference information comprising-category identification information identifying the user, the product sales site, and the category corresponding to the log group, and the category point;

calculating similarity between categories of different product sales sites by applying a second algorithm to the category preference information created for a plurality of users;

creating matching information in which categories of different product sales sites are matched to each other based on the similarity; and

generating recommendation result information comprising a category of a first product sales site matched, based on the matching information, to a category of a second product sales site preferred by a specific user, and providing the recommendation result information to a user terminal of the specific user connecting to the first product sales site.

10 . The method of claim 9 , wherein creating the matching information comprises, for each category of a product sales site, excluding categories of the same product sales site and matching the category to one or more categories of a different product sales site having the highest similarity.

11 . A system for cross-recommendation between online product sales sites, the system comprising:

a plurality of affiliated store servers, each affiliated store server configured to provide a respective product sales site that is a web site to a user terminal of a user, collect behavior log information according to use of the product sales site by the user, and transmit the behavior log information, wherein the plurality of affiliated store servers provide different product sales sites; and

a service providing apparatus comprising one or more processors and memory storing instructions thereon that, when executed by the one or more processors, cause the one or more processors to:

collect the behavior log information of a plurality of users for each of the different product sales sites by communicating with the plurality of affiliated store servers;

group items of the behavior log information into log groups such that each log group comprises behavior log information corresponding to a common user and a common category of a product sales site;

calculate a category point for each log group by applying event types included in the behavior log information of the log group to a first algorithm that assigns a respective weight to each event type;

create, for each log group, category preference information comprising category identification information identifying the user, the product sales site, and the category corresponding to the log group, and the category point;

calculate similarity between categories of different product sales sites by applying a second algorithm to the category preference information created for a plurality of users;

create matching information in which categories of different product sales sites are matched to each other based on the similarity; and

generate recommendation result information comprising a category of a first product sales site matched, based on the matching information, to a category of a second product sales site preferred by a specific user, and provide the recommendation result information to a user terminal of the specific user connecting to the first product sales site.

12 . The system of claim 11 , wherein the instructions further cause the one or more processors to create the matching information by, for each category of a product sales site, excluding categories of the same product sales site and matching the category to one or more categories of a different product sales site having the highest similarity.