Discovering Usage Pattern for Web Recommendation System by using DHP Algorithm

Today, Web becomes a large repository forknowledge discovery. However, how to find neededand related information from the Web is a bigchallenge for users. As a solution, Webpersonalization and recommendation technique haveevolved. Web recommendation is considered as aprocess of identifying user’s preference and adaptingservice to satisfy user’s need based on referring thehistorical behavior of current user or others whoshare similar interest to this user. This paperdescribes how to develop a web recommendationsystem using Web Usage Mining. Firstly, e-librarysystem is developed not only for recording log filebut also for evaluating the effectiveness ofrecommendation system. Then the explanation of weblog data preprocessing is followed. Secondly, how todiscover the usage patterns from Web log files usingDirect Hashing and Pruning (DHP) association rulemining algorithm is described. Finally, rulegeneration processes based on user input ofthreshold values is explained. These rules are used toprovide co-occurrence pages as recommendation touser