Fuzzy-Switch TGARCH Model Applied to Stock Market of NASDAQ Index and Taiwan Weighted Index

• Main Authors: Fang-Yu Hsu, 徐芳瑜
• Other Authors: Yung-Lieh Yang
• Format: Others
• Language: en_US
• Published: 2004
• Online Access: http://ndltd.ncl.edu.tw/handle/77580902017662778358

碩士 === 嶺東技術學院 === 財務金融研究所 === 92 === This paper considers transmissions of volatility in time-varying nonlinear and asymmetric models. Generally, there are many complex factors that can affect transmissions of volatility such as good news and bad news. To account for these elements, we adopt artificial intelligence methodology and propose a Fuzzy-Switch TGARCH model. Specifically, the threshold value in a conventional TGARCH model is modified using rules borrowed from fuzzy logic. Simulations for the NASDAQ index and Taiwan weighted index from January 6, 1992, to June 27, 2002, indicate that transmissions of volatility for the NASDAQ index and Taiwan weighted index are time-varying nonlinear and asymmetric. Furthermore, empirical results demonstrate improved accuracy with a Fuzzy-Switch TGARCH model over the traditional GARCH and TGARCH models.