Application of Time-Series Analysis Methods on Vehicles Sales in Taiwan

碩士 === 中華科技大學 === 經營管理研究所 === 107 === ABSTRACT According to the statistics of the Ministry of Communications, the number of vehicles owned by every 100 people in Taiwan continues to grow from 10.5 in 1988 to 33.9 in 2018 but slow down since 2010. At present, there are more than 7 million vehicles on...

Full description

Bibliographic Details
Main Authors: CHAN LI-FEN, 詹麗芬
Other Authors: LI,BO-JIAN
Format: Others
Language:zh-TW
Published: 2019
Online Access:http://ndltd.ncl.edu.tw/handle/ytp7u5
Description
Summary:碩士 === 中華科技大學 === 經營管理研究所 === 107 === ABSTRACT According to the statistics of the Ministry of Communications, the number of vehicles owned by every 100 people in Taiwan continues to grow from 10.5 in 1988 to 33.9 in 2018 but slow down since 2010. At present, there are more than 7 million vehicles on the roads in Taiwan, but the amazing thing is that more than half of the cars have been on the road for more than 10 years, and there are at least one million vehicles for more than 20 years. In Taiwan, the exchange rate is over 80%. The new car consumer market means that the car is becoming more and more difficult to sell. Taiwan is gradually moving towards an aging society. Like the advanced countries such as the United States and Japan, the age of car drivers is increasing year by year, and the trend of declining children is even more dramatic. The aging of people also affects the enthusiasm for buying new cars. "The car can move and it is still very good," and it is not willing to change cars. The automotive industry has a major impact on the overall economic development. In the forecast of the automobile industry, the rule of thumb is often used as the main basis for forecasting based on past sales data. If it is necessary to objectively predict the sales volume of Taiwan's automobiles, it is necessary to establish an applicable and effective sales forecasting model. In this study, the sales volume of Taiwan from 2012 to 2017 is the training set. The first and second seasons of 2018 are the validation set. The self-regressive model analysis method is included in various techniques of time series. The moving average method, the simple smoothing index method, the HOLT method, the HOLT WINTER method and the ARIMA method are used to establish the forecasting model, and then refer to the overall economic direction to the industrial production index, domestic production gross, investment rate, savings rate, prosperity index, domestic average oil price, Economic growth rate, total income of local residents, excise tax exemption, interest rate, consumer price index, money supply, stock market turnover, stock price index, refund rate, prosperity index, economic response signal and other economic indicators, looking for impact on car sales The larger factor is the independent variable, and the vehicle sales volume is the complex number of strains. In the process of quantitative research, in addition to the traditional regression model, the modern econometric skills are also applied to the study, and the economic indicators are also relevant. Finally, all models are compared and predicted, and the best prediction results can be obtained. KeyWords: Time Series, Regression, Forcasting, Sale of Vehicles