GARCH prediction using spline wavelet support vector machine |
| |
Authors: | Ling-Bing Tang Huan-Ye Sheng Ling-Xiao Tang |
| |
Affiliation: | (1) Department of Computer Science and Engineering, Shanghai Jiao Tong University, 800 Dongchuan Rd., Minhang District, 200240 Shanghai, China;(2) Department of Computer and Electronic Engineering, Hunan Business College, Yuelu Rd., Yuelu District, 410205 Changsha, China;(3) School of Economics, Changsha University of Science and Technology, 45 Chiling Rd., Tianxin District, 410076 Changsha, China |
| |
Abstract: | Volatility forecasting is vital important in finance to reduce risk and take better decisions. This paper proposes a spline
wavelet support vector machine (SWSVM) to forecast the volatility of financial time series based on generalized autoregressive
conditional heteroscedasticity model. An admissible spline wavelet kernel is constructed by incorporating the wavelet technique
and spline theory into support vector machine (SVM). Since spline wavelet function can yield features that describe the stock
time series both at various locations and at varying time granularities, the SWSVM gains the cluster feature of volatility
well. Compared with Gaussian kernel in the standard SVM, the applicability and validity of spline wavelet kernel in SWSVM
are confirmed through computer simulations and experiments on real-world stock data. |
| |
Keywords: | |
本文献已被 SpringerLink 等数据库收录! |
|