Web personalization based on short term navigational behaviour and meta keywords

Abstract

The amounts of information residing on web sites make users' navigation a hard task. To address this problem, web Personalization concept came into existence, which is based on similar users' navigational patterns mined from past visits. Vast techniques are proposed for Web Personalization using Web usage mining but it lacks in recommendation of Web pages which are relative to the user's interest and recommendation of Web pages when the new Web pages which are not in the Web log files are accessed. To solve this problem a novel approach for Web personalization is proposed in this paper which consists of two phases, offline phase and online phase. In offline phase the aggregate usage profile is created by processing the web logs and clustering the sessions obtained from web logs using 'Unweighted Pair Group Using Arithmetic averages' method. And meta keywords of all the URLs present in Web logs are generated which will be used in recommendation process. In online phase the short term navigational behaviour of the user's is used to recommend the related Web pages to the user even though unvisited or new URL is accessed for the first time. The experiment is performed on real data which proved that the system is performing well in recommendation. © 2013 Springer.