| Website is composed of several Web pages,on which the information distributes. Different users get interested in different Web pages.And as the rapidly growing of the network technology,the information in the Internet goes up by index,it's difficult to find the necessary one in so much information.So personalization service seems imperative for us to learn about the use's interests and analyze the browsing patterns by means of Web mining technology,so as to automatically recommend a real-time page for the user and meet the need of conglomeration.By applying the approaches of Data mining into web to solve this problem,a new field "Web mining" is presented.This thesis focus on how to make use of the Web mining analysis log to get the customer's access to the website pattern afterwards put forward a kind of recommendation technique that can give all the customers the sites with the high-efficiency access and the perfect site's topology structure.This thesis mainly has done some research work as follows:1) Summarize the research result concerning the data mining and discuss the current research situation of the data mining based on association rules.2) In order to make the Web log file provide an accurate data source to the mining algorithm,the research carries on preprocessing data and have some problems discussed.3) Introduce the classical Apriori algorithm and FP-growth algorithm,and referring to the process of building FP-Tree in FP-growth algorithm,this paper use a new Homo-FP-growth algorithm to mining the web log data.4) Design a personalization custom-made website system.The research work is a practical improvement to the FP-tree algorithm.It has valuable reference to the research of association rule mining algorithm.The user access the research of mode to be advantageous to raising a site to order information service quality and promote the development of the intelligence information processing,all have important research meaning in both theory and practice. |
PDF Full Text Request