Acquisition of Class Attributes Based on Chinese Online Encyclopedia and Association Rule Mining

摘要

A new method of extracting class attributes based on Chinese online encyclopedia and association rules mining is proposed in this paper. Unstructured texts of encyclopedia articles are taken as the extracting source. Attribute values are viewed as named entities and the frequent pattern mining technique is applied in the proposed method. The candidate attribute words are generated through mining association rules between words and named entities in frequent patterns. Synonymous or duplicate candidate attribute words are found based on the external semantic resource and the computation of similarity of words. Uniform attribute names are obtained by filtering synonymous candidate attribute words. Unstructured texts of six classes are downloaded from Chinese online encyclopedia as experimental data. The experimental results indicate the proposed method is feasible and effective.