Research on web page classification-based core characteristics and web structure

Abstract

The explosive growth of web pages currently makes the research on web page classification technology a hotspot of web mining. This paper introduces experiment data of fashion document corpus by many feature selection and classification methods, gives characterising expressions for specific documents based on core feature terms and web page categorisation algorithm is put forward based on web structure. Through the classification experiment on fashion web pages corpus, the algorithm has higher accuracy rate than other classification algorithms, and thus improves several points relative to the result before adjustment on web structure. The algorithms studied in this paper can be applied in other domains besides web pages of fashions.