Research And Application Of Web Log Mining System Based On Association Rule

Under the circumstance of market economy, the competition in enterprises is to grab customers and markets. That how to improved the strategy and web site structure that satisfied with users to attract more customers and grab mackets is essential to a company. However , traditional way of users registering and logining has not meet the requirements of the enterprise to know the recent and comprehensive information of their customers.With the rapid development of computer and internet,people complete their transaction on web frequently, and abundant data resources in log structure is created on the Web. The data hides users’accessing behavier and their habits.Those enterprise can process the Web data and get the valuable information.In this thesis, on the basis of researching on how to get the user’s information from ths Web logs and referring to the data mining technique,the system architecture of web log minning based on association rule is designed.The layers and functions of the web log minning system is accompanied by illustrations and explanations. Then, the designing flow of system is given and analyzed.By analyzing the web log datas, do the research of data preprocessing and mining algorithm imp- lementing. And association rule way is took in the process of mining algorithm implementing.Based on analysis of the classical charaeter of Apriori algorithm and comparing with other improved algori- thm, an improved Apriori algorithm based on principle of matrix was proposed.Experimental results indicate that,in the same min support and number of log records, the method presented in this paper has better efficiency than other approaches.Finally,this paper designed and implemented the system,which is also being applied to the website of the ship company. The practice indicated that, the appcication of the system make the user’s information gotten by company more exactly and fastly.Meanwhile, it increases the service level for the customers.