An approach of web text information filtering based on domain ontology to expand users' requirements

摘要

Semantic-based information filtering has become a hotspot in text filtering field. Based on the research of traditional text information filtering method which includes too little semantic information to represent user requirements and be understood by the computer, and leads the filtering result not to coincidence of user demands as well as significant influence on filtration performance. This paper presents a web text information filtering method that employs domain ontology to extend customer requirements aiming at these problems. This method includes the semantic extension of user profile based on domain-ontology as well as expression, and the matching algorithm of user demand and the file waiting to be filtered. Firstly, check whether the user requirements belong to domain ontology according to membership degree. Subsequently, deduce and extend the meaning of user demand on the basis of concept description, relational definition and rules of domain ontology. Finally, match document and user demand vector representation after expressing user requirements and file waiting to be filtered by vector space model. The experiment showed this approach can yields good effect on information filtering.