Prediction of car ownership has direct reference significance for the development of urban transportation and construction of urban roads. By analyzing the impact factors of urban auto possession, this paper first analyzes 8 indicators such as urban population, GDP, road passenger traffic and so on determined by some references, then establish BP neural network model to predict the vehicles possession in Hunan Province from 2006 to 2008. The figures of prediction is 989,300, 1,221,800 and 1,370,300 respectively in 2006, 2007 and 2008, which is very close to the real ownership of 946,400,1,217,200 and 1,426,700 respectively. It shows the prediction is very accurate. This suggests that the BP neural network has very strong learning and generalization ability and can be employed in prediction of vehicle possession effectively. The prediction of car ownership, as a foundational work for transportation planning,has direct reference significance on the development of urban traffic,its control and management and construction of urban road, etc.Early in 1940s this research has been started in foreign countries[1]. Many different models of prediction of car ownership have been developed.Many of them are developed mainly based on the factors such as urban economy, population network capacity, the land utilization and parking facilities.In China there are also some researches on this issue. They predicate the car ownership mainly by time series prediction, regression analysis and fractal theory and entropy method [2~6].However, these methods do not comprehensively describe the complex relationship between car ownership and other factors. The author of this paper chooses some car ownership-related factors and employ principal component method to analyze to obtain the main factors, then tries to find the relationship between BP neural networks and car ownership according to these factors so as to predict the car ownership in Hunan Province form 2006 to 2008, which will be greatly significant to the development of urban transportation, management and construction.
Read full abstract