Integrating semantic tagging with popularity-based page rank for next page prediction

Abstract

In this work, we present a next page prediction method that is based on semantic classification of Web pages supported with Popularity based Page Rank (PPR) technique. As the first step, we use a model that basically uses Web page URLs in order to classify Web pages semantically. By using this semantic information, next page is predicted according to the semantic similarity of Web pages. At this point, we augment the technique with Popularity based Page Rank (PPR) values of each Web page. PPR is a type of Page Rank algorithm that is biased with page visit duration, frequency of page visits and the size of the Web page. The accuracy of the proposed method is tested with a set of experiments in comparison to that of two similar approaches in the literature. © 2013 Springer-Verlag London.