The Research Of Data Analysis And Modeling Method For Mobile Behavior

Usually,people think that the individual’s mobile behavior is arbitrary,accidental and unpredictable.However,with the rapid development of mobile communication technology and GPS positioning technology popularization,it becomes possible to acquire large amounts of mobile trajectory data.Through the analysis of mobile trajectory data,it has been found that human behavior has certain intermittent,purpose and cyclical,human behavior has a unique distribution in temporal,spatial and social dimensions.Mobile location recommendation is a hot topic of current research,which is of great significance for mobile network to reduce redundant information and enhance user quality of service.This paper first introduces the related methods and research status of existing mobile location recommendation,analysis the location recommendation model of User-based collaborative filtering(URM)and the location recommendation model of Social-based collaborative filtering(SRM).For the above recommended model in terms of distance similarity and semantic similarity,we design a location recommendation model of Semantic-Distance-based collaborative filtering(SDCRM),calculate the semantic similarity and distance similarity between locations.By clustering the locations based on semantic and distance dimensions,we can acquire a number of location groups and convert the moving trajectory of User-Location to User-Location-Group.Further,we calculate the similarity of behavior between users and recommend locations for the target user.Finally,we choose the Foursquare check-in dataset from location-based social network and analysis the distribution of user behavior on temporal,spatial and social dimensions,validate the feasibility of SDCRM and the precision of location recommendation in the same length of location recommendation list.