| 摘要 |
<P>PROBLEM TO BE SOLVED: To recommend a commodity which is more likely to be purchased by a user even when the commodity has never been purchased in an E-commerce service. <P>SOLUTION: By using information A in purchase information showing whether each user has purchased each commodity and meta-data in commodity information, topics of each user and each commodity are estimated. By using the information A and meta-data B given to each commodity in the commodity information, a relevancy to the purchase information of each word is estimated. By using information in the purchase information showing whether each user has purchased each commodity and the meta-data B, topics of each word and each commodity are estimated. By using user topics, the relevancy to the purchase information of each word and word topics, a probability that each user purchases each commodity is calculated. When a user identifier is given, a recommendation score for each commodity is acquired and identifiers of N top commodities in recommendation scores are selected as recommended commodities. <P>COPYRIGHT: (C)2013,JPO&INPIT |
