Research On Mining Methods Of Following Patterns For Moving Objects

With the help of various positioning tools,a huge amount of mobility data was captured from GPS appliances,mobile phones and wireless networking devices.These mobility data serve as an important foundation for understanding these behavior of moving objects and contain the important information of moving object in space and time.The research result of patterns mining not only helps to understand the behavior of moving objects,but also applies to traffic management,abnormal behavior analysis of animals,path planning and other fields.At present,the moving objects patterns mining mainly concentrate on periodic patterns mining,frequent patterns mining,flock patterns mining,and group patterns mining and so on.In this paper,we will focus on the following patterns mining method of moving object.Intuitively this definition states of following pattern that one entity is following another if it reaches approximately the same location but slightly later.The study of the following pattern of moving objects is beneficial to many practical applications,such as studying the closeness of each moving object in the animal group and tracking the criminals by analyzing suspicious trajectories.The main research achievements are as follows:(1)Due to the existing following patterns mining methods did not take into account the problem of a low sampling rate leads to uncertainty of sampling data,this paper proposes a probabilistic approach and a following patterns mining method based on the Brown-Bridge model.A problem with the device may result in a low sampling rate of the data or inconsistent sampling intervals.In the analysis of these data may result in inaccurate or cannot accurately reflect the real situation.Using the Brownian Bridge to model trajectory of each moving object,then calculates the probability distribution of following.The experimental result shows that the method can mine the following patterns more accurately.(2)Based on the analysis of the characteristics of leading patterns for moving objects,this paper proposes a leading pattern mining algorithm of moving objects based on PageRank algorithm.Because of the similarity between topological structure of web pages and the structure of leading patterns,the PageRank algorithm could be used to solve the problem of mining leadership pattern.The proposed algorithm improves the calculation of PageRank and makes it more suitable for the mining of leading patterns.Because the definition of REMO in the leading patterns is more stringent,the accuracy of the experimental results is not high.The algorithm proposed in this paper solves this problem.According to the experiment,it is proved that the algorithm is effective in mining leadership pattern,and at the same time,it is more accurate compared with other algorithms.