Multiple Computation Models for the Prediction of Private Vehicle Ownership in Chinese Area

Abstract

With the development of Chinese economics, there is an increasing tendency that people have the abilities to purchase private vehicles. However, the increase of private vehicles rapidly increases the air pollution. Therefore, it is necessary that we should study the change regulation of the private vehicle ownership in Chinese area. Due to the impact factors of the change regulation of the private vehicle ownership is various and uncertain, we should take various factors into consideration as much as possible. In this article, we take 11 indicators as the independent variables, while the private vehicle ownership as the dependent variable. Multiple linear regression (MLR) model and artificial neural networks (ANNs) models were developed respectively in order to predict the private vehicle ownership. As an alternative model, we developed grey model GM (1, 1) according to the regulation of the private vehicle ownership. Detailed comparisons among the MLR, ANNs and GM (1, 1) models were made and results show that they have different advantages respectively. Our results shows that all the three types of computation models can be used for the prediction of the private vehicle ownership in Chinese area.