Research on Web log data mining model
Lei Xing-hui · Computer Integrated Manufacturing Systems · 2005
By assigning each webpage a unique number, extracting session from Web log, calculating webpage similarity matrix,and getting user access probability vectors through clustering analysis, the web mining model could then be obtained by using the webpage similarity matrix to modify the probability vectors. The mining result was an integration of page similarity and user clustering profiles. Experimental data indicated that the model could better reflect user access characteristics.