The automatic acquisition of product knowledge in recommender systems

Knowledge based recommender systems can overcome the shortcomings of the traditional ones. To realize knowledge based recommender systems user requirements and resources on web such as information, products and services need to be described as knowledge uniformly. This paper explores the semantic description of products and the automatic acquisition of product knowledge from heterogeneous. databases. The product is modeled as a duple < name, features > and is defined as an ontology using OWL. To acquire product knowledge from heterogeneous databases, we adopt the method of schema mapping and OWL transferring template. The database schema is mapped to XML schema to which the content of the XML documents generated from the data in the databases must conform. The XML documents are further transferred to OWL documents by the OWL transferring template. The generated OWL documents are the semantic descriptions of product instances and can be,provided as input to reasoner such as Racer to conduct ontology based reasoning or rule based reasoning. An example of product knowledge acquisition process is given in this paper.