Abstract:

With the development of cloud computing, big data and mobile internet, mobile user data shows a trend of large data, big noise, increasing dynamic and uncertainty. This impacts the accuracy and efficiency of mobile user data clustering. As a result, we propose an improved custering using representatives (CRUE) algorithm to solve this problem. This algorithm converts the sampling method in the original algorithm, and uses the Map Reduce function to achieve parallel processing. In addition, an interval is used to represent the mobile user data by combining  the concept of interval number. By calculating its interval distance to accommodate the uncertainty of mobile user data, the efficiency and accuracy of clustering are  thereby improved. Finally this algorithm is applied on MIT Reality Project data set, and simulation results  demonstrate the effectiveness and feasibility of the proposed algorithm. It provides support for the further use of mobile enduser data and user's personalized recommendation.

Key words: clustering using representatives(CURE);uncertain data;mobile enduser data;Map Reduce