System and method for social data mining that learns from a dynamic taxonomy

Abstract

A system, method and article of manufacture that selects at least keyword from a taxonomy of a publication system; used the at least one keyword for performing data mining of a social network to detect conversations relating to products; compared elements of the conversations with functions of the publication system; and responsive to comparing, uses the elements of the conversation to change the taxonomy or the offering of products on the publication system. The conversations comprise posts relating to a product, the functions of the publication system comprise searches relating to the product, and changing the taxonomy comprises adding a category to the taxonomy.

1 . A publication system comprising:
 at least one hardware processor configured to include:   a selection module to select one or more keywords;   a data mining module to use the one or more keywords for performing data mining of user communications on an additional system to detect user conversations relating to a plurality of products;   a collection module to compare a number of user conversations on the additional system that relate to a specific product of the plurality of products with a number of searches of the publication system for the specific product;   a detection module to detect that the number of user conversations that relate to the specific product is substantially higher than the number of searches of the publication system for the specific product; and   a database module responsive to the detecting, to change the taxonomy of the publication system and to offer the specific product on the publication system.   
     
     
         2 . The system of  claim 1  wherein changing the taxonomy comprises adding a new category to the taxonomy. 
     
     
         3 . The system of  claim 2 , wherein the selection module selects the one or more keywords from the taxonomy after the adding of a new category or a new product to the taxonomy. 
     
     
         4 . The system of  claim 1  wherein the additional system is one of a social network system or another publication system. 
     
     
         5 . The system of  claim 4  wherein detecting number of user conversations that relate to the specific product comprises detecting conversations of users of the social network that illustrate an interest to obtain the specific product or detecting conversations of users of the social network that illustrate an interest to sell the specific product, the database module further to use the conversations of users that illustrate an interest to obtain the specific product or the conversations of users that illustrate an interest to sell the specific product to revise the way products are offered on the publication system. 
     
     
         6 . The system of  claim 5  wherein using the conversations of users that illustrate an interest to obtain the specific product comprises featuring the specific product on the publication system. 
     
     
         7 . The system of  claim 1  wherein the one or more keywords comprise a plurality of keywords that defines buyers, sellers and products dynamically. 
     
     
         8 . A computer implemented method comprising:
 selecting one or more keywords;   using the one or more keywords for performing data mining of user communications on a first system to detect user conversations relating to a plurality of products;   comparing, by at least one computer processor, a number of data mined user conversations on the first system that relate to a specific product of the plurality of products with a number of searches of a publication system for the specific product; and   responsive to the comparing, detecting that the number of user conversations that relate to the specific product is substantially higher than the number of searches of the publication system for the specific product, and changing the taxonomy of the publication system and offering the specific product on the publication system.   
     
     
         9 . The method of  claim 8  wherein changing the taxonomy comprises adding a new category to the taxonomy. 
     
     
         10 . The method of  claim 9 , wherein the selection module selects the one or more keywords from the taxonomy after the adding of a new category or a new product to the taxonomy. 
     
     
         11 . The method of  claim 9  wherein adding the new category to the taxonomy comprises:
 defining a category of the taxonomy of the publication system as set A; 
 defining a subset of set A as A 1  A 2  . . . An; 
 defining set B as a category based on a social conversation being mined from a social network; 
 defining subset B 1  B 2  . . . Bn as keywords people use to define the category on the social conversation; 
 defining A=B if and only if (A 1  U A 2  UA 3  . . . UA n )=(B 1  UB 2  UB 3  . . . UB n ); 
 detecting B x  as a subset of B, where B x , does not belong to A; and 
 modifying the taxonomy of the publication system dynamically with the data mining by re-defining A as A*
 where A*=AUB x =(A 1  UA 2  UA 3  . . . UA n  UB x ) and 
 A 1  UA 2  UA 3  . . . UA n  is the publication system taxonomy, and 
 B x  is added to the publication system taxonomy. 
 
 
     
     
         12 . The method of  claim 8  wherein the first system is one of a social network system or another publication system. 
     
     
         13 . The method of  claim 11  wherein detecting the number of user conversations relating to a plurality of products comprises detecting conversations of users of the social network that illustrate an interest to obtain a product or detecting conversations of users of the social network that illustrate an interest to sell a product, the method further comprising using the conversations of users that illustrate an interest to obtain a product or the conversations of users that illustrate an interest to sell a product to revise the way products are offered on the publication system. 
     
     
         14 . The method of  claim 13  wherein the publication system is an ecommerce system and revising the way products are offered comprises featuring a product for sale. 
     
     
         15 . A machine-readable storage device having embedded therein a set of instructions which, when executed by a machine, causes execution of operations comprising:
 selecting one or more keywords;   using the one or more keywords for performing data mining of user communications on a first system to detect user conversations relating to a plurality of products;   comparing, by at least one computer processor, a number of data mined user conversations on the first system that relate to a specific product of the plurality of products with a number of searches of a publication system for the specific product; and   responsive to the comparing, detecting that the number of user conversations that relate to the specific product is substantially higher than the number of searches of the publication system for the specific product, and changing the taxonomy of the publication system and offering the specific product on the publication system.   
     
     
         16 . The computer-readable storage device of  claim 15  wherein changing the taxonomy comprises adding a new category or a new product to the taxonomy. 
     
     
         17 . The computer-readable storage device of  claim 16 , wherein the one or more keywords is selected from the taxonomy after the adding of a new category or a new product to the taxonomy. 
     
     
         18 . The computer-readable storage device of  claim 16  wherein adding the new category to the taxonomy comprises:
 defining a category of the taxonomy of the publication system as set A; 
 defining a subset of set A as A 1  A 2  . . . An; 
 defining set B as a category based on a social conversation being mined from a social network; 
 defining subset B 1  B 2  . . . Bn as keywords people use to define the category on the social conversation; 
 defining A=B if and only if (A 1  U A 2  UA 3  . . . UA n )=(B 1  UB 2  UB 3  . . . UB n ); 
 detecting B x  as a subset of B, where B x , does not belong to A; and 
 modifying the taxonomy of the publication system dynamically with the data mining by re-defining A as A*
 where A*=AUB x =(A 1  UA 2  UA 3  . . . UA n  UB x ) and 
 A 1  UA 2  UA 3  . . . UA n  is the publication system taxonomy, and 
 B x  is added to the publication system taxonomy. 
 
 
     
     
         19 . The computer-readable storage device of  claim 15 , wherein detecting user conversations relating to a plurality of products comprises detecting conversations of users of the social network that illustrate an interest to obtain a product or detecting conversations of users of the social network that illustrate an interest to sell a product, the method further comprising using the conversations of users that illustrate an interest to obtain a product or the conversations of users that illustrate an interest to sell a product to revise the way products are offered on the publication system. 
     
     
         20 . The computer-readable storage device of  claim 19  wherein revising the way products are offered comprises featuring a product for sale.