摘要
研究证券市场预测中的股票价格预测精度问题,股票价格受到政治、经济、投资者心理等多种因素影响,股票价格波动较大,系统具有非线性复杂变化规律,单一预测模型只能反映股票价格变化时段信息,预测精度比较低。为了提高股票价格预测精度,提出一种组合模型的股票价格预测方法。首先分别采用ARIMA、GM、RBF神经网络对股票价格进行预测,然后通过权重值获得最优组合预测模型进行股票价格预测。结果表明,组合预测模型提高了股票价格预测精度,降低了预测误差,克服了单一预测模型在股票价格预测中的缺陷,为股票价格等非线性系统准确性预测提供了参考依据。
Research stock price forecast accuracy problem. Stock prices are affected by politics, economy, investors factors and its change rule is complicated and nonlinear, therefore, single forecasting model can not reflect the stock price changes information in short time and the forecast precision is low. In order to improve the prediction precision of stock price, this paper put forward a stock price combination prediction method. Firstly, ARIMA, GM and RBF neural network were used to predict the stock price, and then the stock price prediction resuhs were obtained by the weight value of the optimal combination forecast model. The simulation results show that the combined forecasting model can improve the stock price prediction accuracy, reduce the prediction error, and overcome the single forecast model defects in the stock price forecast. It provides new ideas for nonlinear system.
出处
《计算机仿真》
CSCD
北大核心
2012年第1期356-359,共4页
Computer Simulation