Research And Design Of Personalized Recommendation System Based On LBS

In order to improve the living quality of community residents, and promote the process of community innovative management, community management system based on LBS has been developed in one community project, in which the functions including community information dissemination and public management were realized. With the time passing and business coverage expanding, the information overload phenomenon appeared because more and more information came into the system. The community information service in the current system is still based on simple information display. Mobile Internet technology and data mining technology did not play their due roles in the field of community information.This paper is trying to improve the dissemination of community information by designing and developing the recommendation system of personalized information system based on LBS. The system integrates all kinds of community information, using the localization characteristics of community information to make personalized recommendation according to the needs of individual users.The main research work in this paper is summarized as follows:1. According to the research of behaving data of community residents and community information, we analyze characteristic of community information. Based on the definition and related concepts of personalized recommendation, we definite the objects of community information recommendation system.2. Choose user-based collaborative filtering as the recommendation algorithm, calculate user similarity using the cosine similarity, choose fixed N value of neighborhood selection, predict users’ interest degree of information, and obtain the final recommendation list.3. On the basis of the localization characteristic of community residents and community information, propose an improved collaborative filtering algorithm using data pre filtering. Use hierarchical clustering algorithm to group users in advance, which uses the Geohash algorithm to transcode the location information provided by LBS. According to the grouping results, pre-filter the data to solve the cold-start problem.4. Design the experiment method to verify the effectiveness of the improved algorithm. Implement the community information recommendation system based on the algorithm above in one community project to solve the problem of community information recommendation.