Web-page topical feature extraction for Web-page classification
面向分类的网页主题特征提取*

This paper presented a method that identifies the topical correlativity of one node based on the spatial features, vi-sual features and content features of the page, quantitatively described the different degree of importance of the content, and extracted the topical features through the hybrid weighting method. Experimental results show that Web-page classification based on the extracted page features has better effect compared to the traditional FullDoc text classification.