本文旨在利用 Huang et al. (1998) 所提出的一經驗法則－希伯特－黃轉換 (Hilbert-Huang transform，HHT) 分析方法－對一時間區間內之股價結構轉換 (regime switch) 的次數做一判斷。一般熟知經濟體制顯示其偶爾會從一個結構 (regime) 跳到另一個。一個國家的經濟，週期性地從膨脹轉換 (switch) 到緊縮而後再轉換回去。而一金融市場也會週期性地從低波動 (low-volatility) 轉換到高波動 (high-volatility) 的結構而後再轉換回去。模擬的結果發現，變動的形式為波動變動 (volatility shift) 時，其判斷轉換的時間點較趨勢變動 (drift shift) 來的精準；實證的部分採用臺指 50，並以美國次級房屋信貸風暴為時間區間，對此段時間內判斷出有四次結構轉換。 The purpose of this thesis is to utilize an empirical rule, Hilbert-Huang transform (HHT) proposed by Huang et al. (1998), to determine numbers of regime switching on the stock in a time interval. It is well known that economic systems exhibit occasional jumps from one regime to another. A nation’s economy periodically switches from expansion to contraction and back again, and the dynamics differ between these two regimes. A financial market periodically switches from a low-volatility regime to a high-volatility regime and back again. Simulation result shows that the time period of switching of volatility shift determined by HHT is more accurate than the time period of switching of drift shift. Empirically, we adopt Taiwan 50 index at the event of the U.S. subprime mortgage, and we get that there are four regime switching in this time period.
