An adaptive page clustering based weighting method for information retrieval

With the coming of the era of information explosion, using Internet to obtain information has become the most convenient pipeline information flow. However, the found information mostly based on keyword matching through the search engines, and the search engines do not generally conduct filtering and screening in order to enhance the returns. If the web pages pass a systematic arrangement and are divided into multiple categories or clusters, the users will be guided to obtain real help of information. In this paper, we propose an adaptive web pages clustering algorithm to perform this task. It extracts features to reduce feature dimensions, then filters automatically web pages into its appropriate cluster and enhances the features of the pages to site features for different coefficients to improve the effect. Finally, providing users a more accurate search data model. The experimental results show that compared to the traditional TF-IDF, the proposed approach can find the needed web pages and the topics of the web pages in the corresponding cluster that are highly similar.