A Study of the Forecasting Models and Predictive Factors for Taiwan's Stock Index in Plastics Industry

• Main Authors: Li, Sin, 李欣
• Other Authors: Hou, Chia-Ding
• Format: Others
• Language: zh-TW
• Published: 2018
• Online Access: http://ndltd.ncl.edu.tw/handle/j7s84u

碩士 === 輔仁大學 === 統計資訊學系應用統計碩士在職專班 === 106 === Plastic Industry Stock Price Index in Taiwan always draws a great deal attention in both academia and industry. With correct and proper prediction, it’s possible to grasp the trend of Stock Price Index. Therefore, the study focuses on identifying the major variables affecting Taiwanese Plastics Index, and tries to establish a data analysis model suitable for predicting. The research applies the linear regression models and logistic regression models to analyze 9 variables. And if we take a closer look at each model, we could find out that there are 3 significant variables, which are Monitoring Indicator, Monetary Aggregates-M2, NTD/USD, in the former one; and only 1 variable, Inter-bank Call Loan Rate, in the later one. The coefficient of determination linear regression (R square) is 0.62, and the logistic regression model can obtain in sensitivity of 95%, in specificity of 26.7%, and with total predictive accuracy of 60.8% by a classification point of 0.5. In the time series analysis model, the mean absolute percentage errors of the two best models are 5.68% and 5.67% respectively, which are both much lower than 10%. Thus, it is proved that the models have high predictability.