The Detection Method Of User Check-in Locations Spoofing In Social Networks

With the rapid development of location-based social networks(LBSN),these check-in data from location-based social networks can be used to discover new points of interest and predict user movement patterns,etc.However,since users can intentionally spoof their location when check-in on social networks to obtain expected benefits,there are large discrepancies between check-in data and actual user mobility.Detecting the user's spoofing check-in data can improve the reliability of the social network location service and the credibility of the check-in data.The existing methods for detecting spoofing check-in locations on social networks generally analyze the user's check-in trajectory directly.However,the detection method cannot effectively detect the spoofing location,due to the high sparsity of the check-in data.In this dissertation,we study the problem of user check-in location spoofing detection in social networks.Firstly,to improve the detection accuracy of spoofing check-in data,a text-based location spoofing detection method for LBSN is proposed in this dissertation.The method uses LDA(Latent Dirichlet Allocation)model analysis the historical text information generated by the user and the location text information in the Point-Of-Interest(POI),and the JS distance is used to calculate the similarity between the user and the location text information in the candidate POI,then the probability of the user visiting the POI is obtained to speculate on the probability of check-in location data spoofing in social networks.The results of the evaluation experiments on the real data set show that the method improves the accuracy and recall rate by 14.4%and 10.9%.respectively,compared to the detection method based on the check-in history.Secondly,in order to solve the problem of high sparsity of the user's check-in data,a de-tection method based on text content and joint check-in data is proposed.The method uses the Bayesian model to analyze the user's check-in behavior,and derives the probability of the user visiting the POI,then combines the probability of the user's visiting POI calculated by the sim-ilarity of the text content to infer the probability of check-in data is spoofing in social network.The evaluation results show that the detection accuracy and recall rate of the method can reach more than 60%.In summary,the method proposed in this dissertation can effectively improve the accuracy of spoofing check-in data detection.The research results of this dissertation will help to promote the credibility of location-based social networks and the reliability of spatial-temporal data.