Research And Implementation Of Recommendation System Based On Mobile Environment

This is an information explosion times, it’s very difficult and time-consuming for people to find their own interest or useful information in the face of the vast information on the network. In order to solve the problem of information overload, all kinds of recommendation systems emerge as the times require. With the development of mobile internet and mobile device, because of the advantage of getting information anytime and anyplace, mobile device has replaced computer to be the main platform of abtaining information. Since the mobile device is subject to the small screen, signal instability, poor processing ability and other factors, mobile network compared to traditional network will be faced with more severe information overload problem. In recent years, the personalized recommendation in mobile environment has become a new research hotspot.This paper first introduces the current research in respect of the recommendation system based on mobile enviro nment and some problems existing in current research. Then considering the new characteristics of the recommendation in mobile environment, we use intelligent identification of traffic mode to determine the radius of recommendation, making the candidate items more reasonable.Then we construct user’s interest model dynamically according to the user’s neighborhood and random forest model. After introducing the social network, it can shorten the time-consuming of similarity calculation greatly and enhance the confidence of the user for the recommendation results. Using random forest model can solve the "new item" and "dynamic interest" issues in recommendation system. Based on the proposed algorithm and android developing technology, a personalized recommendation APP based on mobile environment is designed and implemented.Finally, this system has been tested on functionality and performance. According to analysis of test data, the implementation of recommendation system based on mobile environment has good performance on recommendation accuracy and response time.