Indian Journal of Science and Technology

Abstract

Objectives: The small and medium enterprise business utilizes the social networking sites to send advertisement without knowing the user location and information. In this research work, high scalable Location Based-Social Network Advertisement System (LB-SNAS) is developed to find user mobility patterns and provide value service with respect to their geo-presence. Methods/Statistical Analysis: In order to find the user mobility, the continuous geo-data of every user along with the time stamp should be known. In LB-SNAS, a novel Maximum Residing Point (MRP) algorithm is implemented to predict where the user resides most. The spatial and temporal presence of the twitter users is understood by visualizing their geo-tagged data. Findings: Hence, it is observed that 60% of similar user resides at different hotspot at various time periods a day. The existing time based clustering algorithm clusters the user with respect to the time and geo-presence and shows high time complexity. Because whole dataset is subjected to the algorithm. But in the MRP algorithm the difference of each data points are calculated and the data points that are below the threshold are clustered. A single random data point from each cluster is chosen for calculating maximum residing point of user. The execution time is highly reduced in this approach and some useful patterns like similarity mining and crowd strength is determined with respect to spatial temporal parameters. To demonstrate the LB-SNAS, twitter data is used for mining users and their location and foursquare venue data is used as test data to provide location based advertisement. Experimental results show that the users who are constantly moving will have diverse temporary social geo presence and will be definitely having a permanent halt point. Applications and Improvement: A different user who shares similar movements is grouped and hence it derives a key idea for location based group services. This approach helps the online advertisers, social media marketers, small medium enterprise organization to make location based advertisement.
Keywords: Maximum Residing Point Algorithm, Social Data Analytics, Spatial Data Mining, Location based Services, Twitter Data Mining, User Mobility Behavior