User Behavior Research And Personalized Recommendation Design Based On Web Log Mining

Since the advent of the Internet, people’s daily life has been dramatically changed. Nowadays, with the coming of the information age, the information block era has ceased to exist. Whether for the information producer or the information receiver, all surrounded by the problem that "information overload", especially in effectively getting the information they need. So how to process the mass data we face, and to extract useful information we need become the problems to be solved right nowBased on the data mining theory, this thesis research on a company’s website platform, mainly analyze and process the Web log (server log), and further analyze and research the behavior of the Web user. All the company is working for profit, so my research has strong practical significance. On one hand, it could meet the user’s specific needs to a great extent, and then it could optimize the website structure and page. On the other hand, it could make the system be more friendly to the user, and guarantee the sustainable development of the website.Based on the similarity measure, this thesis puts forward a collaborative personalized web recommendation algorithm which is using entropy and then applies it into the recommendation system of the website platform. At the same time, it also takes the specific business of the company and the individualized demand of the user into account.