- 學位論文
臺灣基金持有人投資行為分析─獨立成分分析法與Cox迴歸模型之應用
Investigating the Fund Holders’ Behavior in Taiwan : The Application of ICA with Cox Regression Model
摘要
近年來由於全球投資環境的震盪不安，造成投資人規避較高風險或波動大的投資標的，也間接促使投資金額門檻較低且變現性高等優點之共同基金成為熱門的投資工具。本研究以臺灣某大型投信公司所提供之散戶(小額)投資人實際基金交易資料進行分析(期間為2005年至2011年11月底)，並應用Cox比例危險模型(Cox Proportional Hazards Model)探討基金投資人持有時間長短與外部環境因素的關係。然而在應用Cox比例危險模型時，變數過多常會產生多重共線性之問題，故本研究提出一新穎的多重共線性解決方法─獨立成分分析(ICA)，希望能藉此消除Cox比例危險模型中的多重共線性的問題。為了驗證獨立成分分析方法(ICA)之有效性，本研究以模擬方式比較傳統最小平方法(OLS)、脊迴歸(RR)以及主成分分析(PCA)等方法之結果。實證結果發現，環境因素中與基金投資人持有時間呈正相關且對於持有時間較有影響之變數為股票漲跌幅；反觀，對於基金投資人持有時間較有影響且呈負相關之環境因素為AUD澳幣匯率。
並列摘要
For the past few years, investors tend to avoid the high-risk and high volatility behavior because of the shock disturb of global investment environment. Therefore, the mutual fund with low price and easy change property become the popular tool. The objective of this research is to apply the Cox Proportional Hazards Model in financial data. When the Cox Proportional Hazards Model was used the multicollinearity problem could happen because too many explanatory variables were included. To solve this problem, we propose a new solution─Independent Component Analysis(ICA) to extract the independent components from the included variables in the fitted Cox Proportional Hazards Model. The results not only show that the proposed approach can solve the multicollinearity problem, but also reveals that Australian Dollars exchange rate and stock change extent affect investors significantly.