Web page information was denoted by one-dimension space information function to present the theoretical analysis of information extraction of Web pages in mathematical form. The theorem of Web page information filtration was deducted after analyzing the process of Web information filtration. And then a novel feature extraction theory of Web pages based on correlative filtration was deducted after analyzing similar features of Web pages. Thus based on the feature extraction theory, a novel adaptive information extraction method with feature learning for Web pages is proposed, through combining ...