二层人工神经网络在股市预测模型中的应用——以上证指数为例

Application of Two-Layer Artificial Neural Network in Stock Market Prediction Model—In Case of Shanghai Composite Index

摘要: 文章在阐述了现行的股价指数预测方法基于线性假设的局限性，针对股价指数变动的非线性和波动性，探讨了一种二层人工神经网络的股价指数预测模型。在Python软件的支持下，以上证综合指数(简称上证指数，代码000001) 1050个连续交易日的收盘价为原始样本数据，对指数收盘价及其增长指数进行短期预测，分别计算样本数据和预测数据的涨跌变动趋势。实证结果表明，文中构建的二层结构的BP神经网络预测上证指数变动趋势的准确率高达0.65。得出结论，当样本量和输入层变量选取恰当、网络学习训练次数足够时，使用BP神经网络预测股价指数变动趋势的结果较为准确。

Abstract: In this paper, after expounding the limitations of the current stock index prediction method based on the linear hypothesis, a new stock index prediction model based on the two-layer artificial neural network is discussed, aiming at the nonlinearity and volatility of the stock index, taking the closing price of Shanghai Composite Index (Shanghai Securities Composite Index, code 000001) for 1050 consecutive trading days as the original sample data, then making a short-term prediction of the index closing price and its growth index with the support of Python. Meanwhile, the fluctuation trend of the sample data and the forecast data is calculated respectively. The empirical results show that the accuracy of the two-layer BP neural network is up to 0.65. It is concluded that when the sample size and input layer variables are properly selected and the network learning and training times are sufficient, the BP neural network is more accurate in predicting the trend of stock price index.