Proposed approach toward efficient web usage mining

Web usage mining is the approach that is concerned with mining the log files which stores visitor´s behaviors. The approach uses different data mining techniques in order to discover the hidden knowledge in these files. One of these techniques is the association rule which is very powerful and useful one. The major drawback of it is the huge numbers of the resulted rules that make it so difficult to be analyzed. To overcome this problem the un-useful rules should be pruned and concentrate only on the meaningful ones that really reflect the gained knowledge from the mining process. This paper proposes a pruning technique that success in distinguishing between the valuable associated rules and the redundant ones. The unpruned rules have proven it´s valuable to the webmaster in enhancement the structure of the site and make user navigation as easy as possible.