Friend Recommendationg Based On Social Network And Location Information

Today, various internet applications, such as blog and RSS, blossom on the Internet in this Web2.0era, with online social network service the most popular and fast-growing among them. Meanwhile, the development and popularization of location technique promotes the popularization of services based on locations. With the fusion of social network service and service based on location one of the tendencies of social network development nowadays, how to perform friend suggestion more efficiently turn into a significant question.Based on the traditional friend suggestion algorithm used in social network and combined with users’historical GPS geography check-in, the concept of Social Location Distance is proposed in this thesis. On geography information similarity calculation, except for same check-ins, encounter pattern, same location pattern and same area pattern are also used to find relative pattern couple in the digging of frequent patterns. Then the similarities between each other can be calculated through the check-ins information of pattern couple at different geography areas of different level by weighting in public geography code tree. In the end, in conformity with the similarities between the users’social distance and geography location, friend suggestion is implemented after potential friend suggestion list is queried through Top-k.The experiment results show that friend suggestion based on social network and geography location information is more accurate than traditional friend suggestion pattern.