Friends And Locations Recommendation Framework In Location Based Social Networking Services

LBSNS (Location Based Social Networking Services) is the combination of the LBS andSNS, it has the both characteristics of LBS and SNS. The emergence of LBSNS provides anew form of social service for people and at the same time it produces many new researchsubjects. In these subjects friends and locations recommendation is an important issue. Manyresearchers study on this problem and get some results, but the existing researches are stillinadequate in many ways, it mainly displays in the data processing inaccurate, user similaritycalculation method simple and some others. In order to solve the above-mentioned problems,this paper utilizes the organizational structure of the existing friends and locationsrecommendation framework in location-based social networking services and innovate somealgorithm, so the improved framework is more perfect. The final framework consists of fourmain parts.1. The algorithm of the friends and locations recommendation in location-based socialnetworking services. The algorithm is a guidance algorithm. The other three parts is based onthis algorithm.2. User position data extraction and processing. This step is to handle the coarse data ofthe users, remove the noise data and get the meaningful position information.3. User similarity calculation strategy. This step is the core procedure of the wholeresearch, there are two branches of the user similarity calculation, the first branch calculatedthe user similarity based on the user’s position information and the second branch calculatedthe user similarity based on the user’s behavior semantic, in this paper we combined these twobranches, this kind of treatment is not only improve the calculation efficiency but alsoimprove the accuracy of recommendation.4. Friends and position recommended method. This step is depended on the result of thefirst two steps.Because the content of our research has a strong application value, so we built a friendsand locations recommendation prototype system based on the result of our theoretical study,this system has a certain amount of functions and more perfect functions will be finished asthe future work. Finally, we present a series of contrast experiments and performance experiments, theseexperiments prove that our studies are correct.