Star Anomaly Pattern Mining Based On City Traffic Travel Network

Now China’s transportation infrastructure is gradually increasing.The rapid development of the internet also brings more possibilities for transportation.In an age of information,faced with the diversification and complexity of transportation methods,the research group constructs an intelligence information analysis framework IAF.IAF perceives,understands and predicts the ground,event,people,behaviors and mutual relationships,which is based on complex network mining algorithms.Anomaly detection is a very important part of IAF.Intelligence analysis based on traffic complex network is an important function of IAF.This paper obtains car-hailing trajectories from GPS data and shared bikes’ travel data in physical space.Based on the anomaly subgraph mining algorithm of a specific structure,we detect anomaly pattern in the city traffic travel network,and do the following work:Firstly,we present a framework for mining star anomaly pattern in city traffic travel network.Through the extraction of city traffic travel data,we build an attribute network with community information.Then according to the characteristics of the city traffic travel network,the framework detects the largest star anomaly subgraph.After that,the framework clusters the largest star anomaly subgraph at various time periods,and defines the analysis process of the network topology.Secondly,we complete cleansing and fusing huge amounts of car-hailing trajectories from GPS data and shared bikes’ travel data in Tianjin,which form a data set of city traffic travel in Tianjin,and acquire community data in Tianjin.After that,we fuse city traffic travel data with the community attribute and overview the data in from multiple views.And we calculate the feature of Tianjin city traffic travel network.Finally,we mine anomaly pattern on Tianjin city traffic travel network,and analyze the mining results.The experimental results show that there are certain rules for abnormal patterns at several time periods.To a certain extent,it is helpful for analyzing and predicting the traffic management and emergencies effectively at peak periods.This paper presents a framework for mining star anomaly pattern of city traffic travel network.Through the processing of Tianjin’s shared bikes and car-hailing data,we build city traffic travel networks with community information,verified the efficiency of the method in the network,explain the reasons for the abnormal patterns,and analyze the similarities and differences of the two modes of transportation.Experiments show that the framework can effectively identify abnormal patterns in city traffic travel data and provide some suggestions for traffic management.