Research Of Recommended Algorithm On Real-time Authentication System

With the development of Internet technology, network traffic and the number of network users has reached an unprecedented scale, people in public places need security Internet environment and personalized shopping. Based on the current research related to the analysis of authentication technology, we design the wireless authentication system that based on web server, and then we bring the websocket real-time communications technology to this authentication system and inplement it; and at last we have combined the real-time authentication system with personal recommendation technology to establish an effective advertising recommendation algorithm, to help businesses to conduct effective marketing.This thesis is divided into two parts, one is the design and implementation of real-time authentication system; and the other part is the analysis and implementation of ad recommendation system that based on real-time authentication systems.This thesis studies the principle of certification based on wifidog and real-time communication based on websocket, which are the important theory in our real-time authentication. We compared websocket with traditional ajax long polling, iframe streamed real-time communication technologies; and then we design and implement the certification system through WAMP development framework; at last, we analysis the principle and differences between traditional web-based authentication and real-time communication technology based on websocket, and then we design and implement real-time authentication function by the way of websocket client and websocket server.Another part of this thesis is about the recommendation of personalized ads for the authenticated user. First, we analyzed the classic content-based recommendation, association rules, collaborative filtering recommendation algorithm principle; and then we show the principle of recommendation algorithm based on real-time authentication system; finally, we have build the recommendation model based on the content-based recommendation algorithm, and combined the improved string matching algorithm with Chinese characters fault-tolerant algorithm, we design and implement the recommendation system. At last, we analyzed the performance of the algorithm through the result.Finally we test the function of registered certification, certification by mobile phone, and no-authentication respectively; and then testing the function of real-time authentication based on websocket successfully; and then, we have test the ads recommendation algorithm based on the real-time authentication system. In the end, we analyzed the performance of the algorithm through the test result, and we have got ideal result.