Applications of HAR-GARCH Models

• Main Authors: Chang, Yi-Cheng, 張益誠
• Other Authors: Huang, Shih-Feng
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/dy2tky

碩士 === 國立高雄大學 === 統計學研究所 === 107 === This study proposes to construct association rules for global economic conditions by fitting hysteretic autoregressive models with GARCH in mean effects, denoted by HAR-GARCH, to financial time series. A Markov Chain Monte Carlo algorithm is employed to estimate the model parameters and the economic conditions of a financial market are obtained accordingly. In the empirical study, we collect 13 global stock market indices from August 1, 2008, to August 30, 2018, and fit HAR-GARCH models for their daily returns. The association rules of the 13 stock markets are established based on the fitted models. In addition, the mutual interactions in the financial markets during the investigation period are better understood through various indicators of the association rules.