FIGARCH model on Chinese securities market based on the genetic algorithms

In this paper, a new method of Fractionally Integrated Generalized Autoregressive Conditionally Heteroskedasticity (FIGARCH) model for characterizing financial market volatility is introduced to test the long memory property. We also introduce a new method to establish FIGARCH model - Genetic Algorithms (GA). The paper makes a program according to GA in C to analyze the integrated index in Shanghai Stock Exchange. We study the errors in AR model and find that they have heteroskedasticity with the software Eviews. Then we establish FIGARCH model with the errors and get the values of the parameters by the program in C. The empirical results illustrate that there is apparently a long memory property in the Chinese stock market volatility.