Research And Implementation Of Web Usage Mining System Model

With the development of Internet, the sheer volume of information available on the Internet is overwhelming. This phenomenon is referred as information overload. The information diversity on the Internet makes it even harder for users to find the desired information. Users are lack of effective ways to find relevant information and get lost easily, namely information bewilderment. Nowadays, we primarily use search engines for information retrieval. Most search engines, however, perform passive searching and return results regardless of the preference or specific interests of different users. Therefore, search engines cannot solve the information overload and information bewilderment problems effectively now.Among many direct and indirect solutions, applying data mining techniques on Web log is a promising approach. The Web log data is generally vast and redundancy. Moreover the relations among Web pages are fuzzy and uncertain. Rough Sets theory is a soft computing tool of dealing with uncertain and incomplete information. Fuzzy Logic is an analysis method by establishing fuzzy analogical matrix based on the character, distance and similarity among objects. Association rule mining is finding interesting rules between itemsets from large data sets. It's one of key technologies in data mining. Web Usage Mining can get interesting pattern from the log data of websites. According to the browsing behavior of users, we can improve the structure of website and provide personalized services for users. So the research of "Web Usage Mining System Model" is of theoretical significance and realistic value.The entire process of data mining, Web mining and Web usage mining is introduced systematically. And an improved Apriori algorithm based on adjacency list and index is proposed. One Web Usage Mining System Model is designed and implemented. The model has been tested with the log records of our campus website, and the result is satisfying. At last, I summarize the advantage and deficiency of the model, and put forward the objective of more research.