This paper aims at studying the way of combining data mining technology with association rules and the way of looking for useful information hiding behind the data from the large amounts of Web-log data. First, the author studied the relevant technique about data stream mining and mining of web browsing patterns; and determined to analyse the browsing patterns of users by the algrithem of Lossy Counting,which is an important technique based on association rules for data stream mining.Secondly, the author analysed and optimized this algrithem with the nend in reality.At last,the author applied this method to find the information of users in a real web log data,the result indicate that this system can work perfectly,and the research is useful to improve the management of web site.
|