Research On Combining Web Content Mining And Web Usage Mining

Web mining attempts to discover the useful knowledge and patterns from the web resource. According to the variety of the web data, web mining is classified as three parts: web content mining (WCM), web structure mining (WSM) and web usage mining (WUM). Each of them has great potential value and broad development prospect, For example, WCM for the pages auto classifying, WSM for the search engine, and WUM for personal recommendation, etc.The former researches in web mining are focus on a certain part. But web data, as an existent entity, has complex relation with each other. Therefore, combining different part of web mining will be a new research field.Based on the considerations above, a new research method of combining the WCM and WUM is introduced by the deficiencies of the two. In the process of clustering the patterns, a method that considers not only the sequence of web pages but also the similarity of pages in the same index of different sequence is proposed. By this way, the accuracy of the result is improved.In computing the similarity of two patterns, the patterns are divided into two parts: the same part and the different part, and are measured respectively, thus avoid the error that patterns are not similar because the sequence are different but the intent are same.Applying the method of combining research, more useful knowledge can be obtained.