Volatility forecast by discrete stochastic optimization and Genetic Algorithms

被引：0

作者：

Ma, I ^{[1
]}

Wong, T ^{[1
]}

Sankar, T ^{[1
]}

Li, L ^{[1
]}

机构：

[1] Univ Quebec, Ecole Technol Super, Dept Automated Mfg Engn, Montreal, PQ H3C 3P8, Canada

来源：

Proceedings of the 8th Joint Conference on Information Sciences, Vols 1-3 | 2005年

关键词：

volatility; financial index; forecasting; stochastic optimization; evolutionary algorithm;

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

The daily volatility is crucial in the study of financial risks. In an earlier attempt, a group of rules is successfully trained by Genetic Algorithms (GA) to extract patterns from the Integrated Volatility (IV) time series enabling analysts to achieve a forecasting accuracy of 75%. The current paper substantiates such a use of GA with a Markov chain based discrete stochastic optimization method. By transforming the IV time series into a Markov chain, it proves the link between the two methods in case of time non-homogeneity and convergence. Viewed differently, it demonstrates the efficiency improvement GA brought to the application of the stochastic optimization method in the forecast of IV.

引用

页码：992 / 996

页数：5