摘要
金融时间序列预测是金融理论领域的研究热点之一。以金融市场中普遍存在的弱混沌为基础,对递归预测器神经网络在中国金融市场的预测应用进行实证研究。在网络训练上,提出用遗传算法优化网络的阈值、权值以及激发函数的幅值和斜率。对国内股票、期货和黄金市场中几个有代表性的品种进行实证检验,计算了预测均方根误差(RMSE)和预测精度(PA),并和两种典型的神经网络预测模型——BP神经网络、径向基函数神经网络做了比较,结果表明该模型有较好的预测效果。
The prediction of financial time series is one of the hot researches in finance.Based on the existence of weak chaos in financial market,we make an empirical study on prediction of Chinese financial market based on the recurrent predictor neural network(RPNN).We train the network by using genetic algorithm(GA) to optimize the weight and threshold,also the amplitude and slope of the excitation function.We test some typical varieties of Chinese stock market,futures market and gold market.Two performance measures(root-mean-square error(RMSE) and prediction accuracy(PA)) are calculated and compared to two classic neural networks-back propagation neural network(BPNN) and radial basis function neural network(RBFNN).The result shows that the proposed method is more effective and accurate.
出处
《统计与信息论坛》
CSSCI
2013年第1期14-21,共8页
Journal of Statistics and Information
基金
国家社会科学基金项目<发展循环经济与中国特色的再制造产业协调发展研究>(10BGL010)
关键词
金融市场
混沌
递归预测器神经网络
BP神经网络
径向基函数神经网络
遗传算法
financial market
chaos
recurrent predictor neural network
back propagation neural network
radial basis function neural network
genetic algorithm