A Data-mining Based Research On The Analysis Technology Of The Cell Phone Communication Users' Behaviors

Data mining is a major research area currently.With the rapid development of information technology,more and more people use cell phones to communicate with each other,so the bill data is generated.These bill data could be applied to locate users,and further to study and analyze the users' travel process as well.Based on the large-scale mobile location data,this thesis studies the traveling crowd behavior.First,the data have been preprocessed and the visitors have been screened.Due to the influences from both the network itself and the environment,large quantities of noisy data have been existing in the cell phone data collected at first.Therefore,these noisy data have been deleted and corrected accordingly.As a result,a set of filtrated rules aimed at travelling crowd has been developed based on the dealt cell phone data.Then,in order to recommend the popular tourist spot to tourists,the commonly used association rule mining algorithm has been comprehensively studied and analyzed in this thesis.Aiming at to solve the problems of the Apriori algorithm,an improved Apriori algorithm based on the longest common subsequence has been put forward.The feasibility of the improved Apriori algorithm has been verified by using the existing location data.Finally,the nearest neighbor algorithm of the Voronoi graph for the target search is improved and an In-Route Nearest Neighbor algorithm in the path is proposed based on the weighted Voronoi graph.In addition,the ArcGIS has been applied as the development tool that helps to realize the division of the map,which has also been proved in the actual data.