Research On The Recommendation Method Of Mobile Network Advertisement Based On Collaborative Filtering

With the rapid development of information technology and the Internet, the number of Internet users is increasing year by year. As of December 2015, the number of China's Internet users has reached 688 million. Among them, 39.51 million users are new Internet users added in this year; the growth rate is 6.1%. At the same time, the ways of getting information have also changed dramatically. In the past, people found information on the Internet, but now,filtered information from the various smart devices. Information overload will become a major problem in the development of the Internet. In the era of explosion of information,advertising industry has being developed rapidly with the Internet. Therefore, we can see many ads around us. As of 2015, the size of Chinese advertising market has reached 209.37 billion. In the next few years, the scale of advertising market will increase with the development of online advertising market. With the widely using of mobile phone,tablet,mobile device and with the appearance of social network and LBS technology,the form of recommend ads becomes diverse and the content is more rich and colorful. Personalized and precise advertising has been payed more attention by every advertising company, and it will become the main trend in the future. In this thesis, we researched the user's interests and activity districts, and proposed a method of ads collaborative filtering recommendation. The main research contents of this thesis:(1) Using Key Words Mode to analyze the user similarity. We dig users' interests through analyzing the Weibo published by users and then divide user interest degree into different levels. At last, compute the similarity between users.(2) Analyzing the users' check-ins by clustering and hierarchy. And then, we calculate the similarity value of two users' activity districts in space and time combined with time slots.(3) Proposing a method of user-based collaborative filtering recommendation for mobile advertising. The user's interest similarity and activity districts are considered into together as a whole. And then, we find “neighborhood” according the similarity between users. At last,we recommend ads among the “neighborhood” by using collaborative filtering with the time and the location of user clicking ads and the popularity of ads.(4) Simulating the process of recommendation by experiment and analyzing the recommendation lists affected by others factors. At last, we evaluate the result of recommendation through experiments.