Clustering system, method, and program, and recommendation system

Abstract

An object is to provide a clustering system capable of performing clustering of a plurality of types of items to be able to recommend an item whose corresponding textual data exists but relational data with another type of item does not exist, to the other type of item. A first clustering meansperforms clustering of first IDs, based on the relational data. A second clustering meansperforms clustering of second IDs, based on the relational data and textual data associated with the second IDs. A topic assignment meansassigns a topic for each word included in textual data corresponding to each second ID. A parameter decision meansdecides a parameter used for first clustering processing, a parameter used for second clustering processing, and a parameter used for topic assignment processing. The processing described above is repeated until it is determined that a predetermined condition is satisfied.