The Research And Application Of Data Mining On The Data Of Bus Intelligent Card

With the development of urbanization and the rapid increase in the number of vehicles, transportation and socio-economic life have become more and more close. Road transport has become the most important mode of transportation in the ground. However, too many vehicles lead to increased traffic congestion, frequent accidents and other issues of aggravating environment and so on. How to improve the traffic capacity of roads to prevent the occurrence of congestion, urban transport planning seems to be more important, and accurate traffic prediction is a key of transport planning issues, so the traffic forecasts are get more and more attention by people.Road transport system includes such as time variability, complexity and uncertainty, which are too difficult to acquiring basic datas on traffic and predicting traffic quantity, and these are difficult reasons to forecast traffic. With the rapid development of information technology, Intelligent Transportation System (Intelligent Transportation System, referred to as ITS) to solve traffic problems in the area show an increasing number of advantages.Rcently, the use of IC cards can provide a large amount of datas to the public traffic. The rising database and data mining technology make the prediction that is both efficient and accurate and decision-making possible.Based on the database and data mining technology and connected with the specific situations of public transportation system in Changchun City, the thesis propose to apply the database and data mining technologies to analysis of data of bus intelligent card. Above all ,this article describes the database and data mining technology and commonly used data mining algorithm; and then studied the database bus intelligent card in the application of data mining.It also introduced the bus intelligent card information, vehicle time information, bus route information, and site information which are public transportation data collective methods. Design concept model, logical model and physical model of the three-tier data model introduced the use of Microsoft Analysis Services for multi-dimensional data modeling process.According to Changchun city's bus intelligent card payment system, the use of SQL Sever 2000 Analysis Services design and build a database bus intelligent card.Traffic analytical indicators of research lines, including: traffic routes of all the day,rush hour traffic, uneven coefficient of roads,directions and variation a day and so on. Based on these assorted indicators,we can bent on making analysis on the characteristic of time distribution of all traffic , of single-route and travel time of different groups .Changchun City bus intelligent card data ,as an xample, give the corresponding distribution in order to analyze the characteristics of public transport residents heterogeneity of time travel , the direction of non-uniformity and the variety of flow-week cycle.We have discussed that it is practicable that artificial neural network apply to traffic prediction in the line, the use of regressive analysis can help to rebuild traffic prediction model which based on BP neural network.There are discusses for the design of input and output layer, identification of network layer, the number of neurons made of various levels, selection of enabled function , initial value and learning rate, as well as the determination of the relevant parameters, finally ,the simulation results of research are given in order to depict the value of practical application of the model .The main research results:1. Database and data mining technology apply to public transportation bus intelligent card system .The use of database and data mining technology to analyze the data bus intelligent card can make significant meaning on the public transport programming and give a new conception which is how to make full use of bus intelligent card data bus for public transport management and public transport programming.2. BP neural network algorithm will be used in forecasting traffic, the predictive model of established BP neural network have a high predictive accuracy and practical significance.