本文选取了沪深300指数和百度股票的收盘价，利用ARIMA模型和BP神经网络两种单一模型以及两种模型的组合对股票价格进行预测，其中组合模型采取了等权重组合和方差倒数法两种定权的方法来确定权数。结果表明，通过等权重组合方式的模型ARIMA-BP的预测精度最高，预测的效果最好，BP神经网络模型效果其次，效果较差的为ARIMA模型。
This article selects the CSI 300 Index and the closing price of Baidu stocks, and uses two single models of ARIMA model and BP neural network and a combination of the two models to predict stock prices. The combination model adopts two weighting methods: equal weight combination and reciprocal variance method to determine the weight. The results show that the ARIMA-BP model with equal weight combination has the highest prediction accuracy and the best prediction effect, followed by the BP neural network model, and the ARIMA model with the poorer effect.
