Eliminate Interference Of Season To Analyse Consumption Of Private Car

Since the year 1990, car consumption in China, has been becoming more and more important position in the economic activity, and auto industry development is of great significance for China's overall macro-economic development.It is predicted that the next 10-15 years, GDP per year of 16%-17% new capacity will be occupied by the automotive industry, car industry will be the strongest driving force of our national economy in the industry. Therefore, to predict the demand for private cars will help enterprises to formulate development plans and the Government's industry development plan in order to protect the health of the national economy sustainable development.Article summarizes the growth model cars at home and abroad in research methods and theories, we decide based on the models of private vehicle growth Gomperta model, log-linear model and the distribution delay model further and use structural time series analysis and multivariate statistical analysis in combination method. We eliminate the seasonal cycle of automobile consumption data of factors such as interference exists, then set up a target hierarchy, the principal component simplified indicators. With strong market demand for private cars explanatory prediction model.This paper analyzes the characteristics and impact of China's auto market consumption of private cars in China, the main factors such as the main macro-economic considerations, market environment and other.these three factors further sub-layer of factors under the total of identified 9 indicators. Then we classified the main components of the index extraction, come to reflect the three factors were the main changes in the five main components. Consumption data for the existence of another vehicle factors such as seasonal cycle of interference characteristic were heteroscedastic testing and ARMA model and simulation, using time series analysis and the variable parameter state space Kalman filter method of thinking of the car sales data for the seasonal factors removed and data correction. Then we use the revised data and principal component scores of data to the Lord to explain the variable component, the market demand as the dependent variable, to explain the established regression prediction model, and fitting the model parameters are tested statistically superior degree and concluded that model significantly. Finally, we conducted a model of stability analysis and some key good comparison of forecasting methods, found that the model has a more stable bond, comprehensive multi-factor, has strong explanatory power of the market, forecast error is small, the model is applicable. In this paper, a combination of research methods used in KALMAN filter time series analysis and multivariate analysis methods.