A new approach for web usage mining using case based reasoning

Abstract

In this study, we present a new approach for Web Usage Mining using Case Based Reasoning. Case-Based Reasoning techniques are a knowledge-based problem-solving approach which is based on the reuse of previous work experience. Thus, the past experience can be deemed as an efficient guide for solving new problems. Web personalization systems which have the capability to adapt the next set of visited pages to individual users according to their interests and navigational behaviors have been proposed. The proposed architecture consists of a number of components, namely, basic log preprocessing, pattern discovery methods (By Case Based Reasoning and peer to peer similarity—Clustering—association rules mining methods), and recommendations. One of the issues considered in this study is that there are no recommendations to those who are different from the existing users in the log file. Also, it is one of the challenges facing the recommendations systems. To deal with this problem, Apriori algorithm was designed individually in order to be utilized in presenting recommendations; in other words, in cases where recommendations may be inadequate, using association rules can enhance the overall system performance recommendations. A new method used in this study is clustering algorithms for Nominal web data. Our evaluations show that the proposed method along with Standard case-classified Log provides more effective recommendations for the users than the Logs with no case classification.