Data Mining And Application Research Of The Vehicle Route Distribution In Provincial Highway System

The Connected to the Internet Charge Data of highway contains a huge number ofvehicle records driving in the road network, which included rich Traffic behaviorinformation.With the different from the traditional method, such as vehicle detector or videodetector, The Connected to the Internet Charge Data is aim at the single target vehicle. itprovided a research approach which from micro to macro or from vehicle to road net.And itcan grasp the features of traffic behavior, traffic structure and traffic distribution in thenetwork accurately. Because of the vehicle detector are prone to break down in the actuallayout. When it comes to vehicle detector as a means of data collection, the data integrity isaffected by the factors such as power supply, network, communication and so on. If videodetection is used as a means of data collection, due to the Large amount of video data, itdemand higher transmission process, which is not conducive to real-time data analysis andprocessing. because of tactual operation requirements, we must ensure that the data security,complete, timely transfer. besides, The Connected to the Internet Charge Data can accuratelyreflect many traffic information, Like traffic type, route and travel time and so on. therefore,as a window on road network running status, The Connected to the Internet Charge Data hasmore obvious advantages, and is more appropriate for the data mining in actual businessapplication.Based on the traffic distribution structure of road network, This article research datamining using in traffic flow predictive. Its data source is come from huge raw data of TheHighway Connected Charge System, combining with the structure of actual trafficdistribution and the application of highway operation management department. BP neuralnetwork algorithm was studied for the short prediction of traffic flow. We can obtain trafficflow estimation at different toll and at different time within short time in the future. In order toguide practical operation coordination and optimal allocation of resources. At the same time,guidance timely in the time of special period or the traffic peak period will be helpfull forreducing traffic delays, and improve road network efficiency, or promote the level of thehighway operation management and service. The research in traffic flow prediction which based on traffic distribution structure a integrated embodiment, which apply data miningresults in the actual business.At the same time, The data mining research which based on traffic distribution structurecan analysis the stability of network traffic distribution structure. After verification, there istraffic distribution stability between two toll in the actual network operation. the week trafficflow of different tolls has the characteristic model. by means of analysis traffic behavior ofspatial flow patterns between cities in the area, we can discuss the association rules betweentraffic behavior and urban economic activity in terms of seasons, cycle and route, whatcreated a new thinking of transportation systems interact with the social and economicresearch, and realize the traffic data mining results applied in the field of social economy.