Study On Location-Based Personalized Recommendation System In Mobile Advertising

Mobile Internet is developing rapidly in recent years, access to Internet by mobile devices has increased significantly. At the same time, the Location Based services will show a flourishing trend. With the popularization of smart phones and mobile devices, mobile advertising market develops fast, for the new market has enormous potential, many companies have placed concern in this area. Currently, mobile advertisers use traditional methods of ad targeting. But in five years, mobile ad targeting will become more relevant to a person’s behavior and current location. However, users are still facing the problem of information overload. The limits of a user’s time, energy, environment, screen of mobile terminals will make this problem even worse.In response to this situation, the personalized recommendation system for location-based mobile advertising is studied in this paper. The purpose of the system is to provide location-triggered real-time search and location-based, situational/interest-triggered real-time advertising information referral service for mobile users. This paper studies two aspects. On one hand, personalized recommendation technology is studied to solve the problem of information overload; traditional collaborative filtering recommendation system applies the Matrix factorization technology to solve the problem of data sparseness, but the social relationships between the users is not paid enough attention. This paper studies the merging of Social Relationships and Collaborative Filtering, and provides a new attempt to improve the recommendation accuracy. On the other hand, this article builds the geographically partial matching algorithm based on the proxy server architecture; the algorithm will compare the cache data with the advertising inquiries submitted by mobile users, if the similarity of the search area reaches a certain level, cache data will be transferred back to the user as a result; thus greatly enhance the reaction efficiency of the system, bringing a better user experience. Finally, simulation results show that the geographically partial matching algorithm can significantly increase the cache hit rate, the test results prove that the algorithm has a higher value in real-world applications.