Sales Forecast Of Automotive Market And A Discrete Choice Model

The dissertation discussed several issues from different aspects of automotive market, including auto sales forecast, market shares, together with the effects of gasoline price, tax policy and pricing strategy to the market. Integrated with actual market data, these factors were modlized and derivated some conclusion or discussion according to different topics.We started from a regional market in Chapter 1, an analysis focused on the sales volume of a fix model was performed, moreover, utilizing PCA and Time Series approach, a sales forecast model was obtained. The forecast performance of the model was tested by real data.Chapter 2 through Chapter 4 were emphasized on the discussion of a whole market and were related to each other. Chapter 2 began with a Random Choice Model. Combined with random utility model and several market and product factors, a consumer choice model was developed in this chapter. A systematical discussion of the influence of gasoline price to the market sales of vehicles with different gas economics was then processed. Based on the conclusion of Chapter 2, Chapter 3 discussed how the new sales tax policy in China would affect the Chinese automotive market and the oriental strategy of the auto manufactories. Chapter 4 further derived the random choice model developed in Chapter 2, obtained some results on the issues of maximum utility in a market, new member of a market and the interference of consumer attributes, as well as an explanation of the economical sense of these results.The content in Chapter 5 and Chapter 6 is more theoretical. Chapter 5 included some works finished several years ago when I was major in Statistic. Utilizing some approach fo Limit Theory, this chapter mainly discussed some asymptotic properties of the sum of record values and some of the conclution was applied in Chapter 6 to analyze the limited property of Random Utility, also in Chapter 6 a discussion of the limited property of choice probabiliy of Discrete Choice Model was performaned, this study is mainly based on the tools of Regularly Varying Function.