Structure-Based Web Pages Clustering

Recognizing similarities among the documents of a set is one of the objectives of retrieving information. The information related to the similarities of web pages can be used to present similar documents to users in order to retrieve considered information. In the present study, a new algorithm has been proposed to cluster web pages based on their structure. The proposed algorithm is based on hierarchical clustering designed based on SCAN algorithm. Cross-page link and in-page link structures have been tested to create a new similarity function and clustering algorithm based on volume criteria has been created for web pages in order to determine hierarchical relations. As the research findings reveal, the proposed method is effective to be used in uncovered structures of web pages in some organizations` websites.