近年來風險管理成為重要的議題，而風險值是一個測量市場風險的 指標。本篇論文採用許多方法來計算風險值包含無任何假設方法以 及假設方法。無任何假設方法為歷史模擬法，過濾歷史模擬法以及 RiskMetrics。假設方法為 GARCH-normal 模型，以及 GARCH-t 模 型。為了確定計算風險值方法的準確性，我們利用 Kupiec 的 POF Test，Kupiec 的 TUFF Test，獨立性覆蓋檢定，以及條件下的覆蓋 檢定。我們主要是探討哪一個計算風險值的方法是較準確的。;Recently, risk management has become an important issue, and value at risk (VaR) is a index to measure the market risk. The thesis adopts several methods to calculate VaR, including the non assumption methods and assumption methods. Non assumption methods like historical simulation, the filtered historical simulation, and the RiskMetrics method. Assumption method like the GARCH-normal models and GARCH-t models. In order to check the accuracy of the VaR calculation methods. We consider Kupiec′s POF Coverage Test, Kupiec′s TUFF Coverage Test, Independent Coverage Test and Conditional Coverage Test. We focus to check which of the VaR calculating methods is more accurate.
