- 學位論文
應用LASSO迴歸探討台灣景氣領先指標、總體經濟指標與股價關係
Applying LASSO Regression to Discuss the Relationship among Taiwan Leading Indicators, Macroeconomic Indicators and Stock Variables
摘要
景氣是影響股價的因素，而行政院經建會所編製的台灣景氣指標中的領先指標，具領先於景氣變動的特性，常被用來探討與股價的關係，但此類總體經濟指標於迴歸分析中，易產生多重共線性的問題。多重共線性是迴歸分析中的自變數間存在相關的現象。嚴重時將對迴歸分析造成影響。 LASSO迴歸是ㄧ種以降低多重共線性為本質且同時具備自動變數選擇的迴歸方法，在國外已有許多案例將其應用於經濟或其它領域的分析，可在台灣卻少有相關應用。 因此，本研究之主要目的在於應用LASSO迴歸，分析包含台灣景氣領先指標在內的總體經濟指標與股價之間的關係，由LASSO迴歸的角度找出影響股價的因素，並舉個股為例，提供投資人簡單的投資參考。
並列摘要
The boom is one of the factors that affect the stock price, and leading indicators of Taiwan Business Indicators compiled by the Executive Yuan is often used to explore the relationship with the stock price, but these macroeconomic indicators are easy to produce multicollinearity in the regression analysis. Multicollinearity is the existence of correlation between the independent variables in the regression analysis. If this situation is too serious, it will have potential adverse effects on the regression analysis. LASSO regression is a kind of regression method that can reduce the multicollinearity and has the ability of automatic variable selection. There are many cases abroad that will be applied to economic or other areas of analysis, but in Taiwan, there are few related applications. Therefore, the main purpose of this study is to use LASSO regression to analyze the relationship among Taiwan leading indicators, macroeconomic indicators and stock price. Then, find out the factors that affect the stock price for investors as a simple investment reference.