基于GA-Elman动态回归神经网络的股价预测模型研究

Stock Market Forecasting Research Based on GA-Elman Neural Network

文章针对股价预测问题的复杂性、不确定性、时变性及动态性等特性,利用Elman神经网络具有记忆性的优点,采用遗传算法训练优化Elman神经网络的初始权值,提出了高效的GA-Elman动态回归神经网络股价预测模型。实验模拟结果表明:该模型快速稳定且具有较高的精度,将其用于股票价格预测可行且有效。该模型的提出也为股票价格预测提供了一种新的技术和方法。

Based on characteristics of stock prices and the merits of Elman networks, this paper establishes a GA - Elman neural network model, in which Genetic Algorithm is used to optimize the weight values. So the model efficiently solves the problems of complexity, un- certainty and dynamic characteristics of stock prices prediction. Simulation results show that the proposed GA - Elman neural network is feasible and efficient. This paper provides a new method for modeling and forecasting in stock market.