基于灰色系统和神经网络的创业板股票价格预测研究

股票价格预测是投资领域的一个重点关注课题。由于股票价格受到诸多非线性因 素的影响，得到精确的预测结果较为困难。为了消除股票指标的多重共线性，采用Adaptive- Lasso算法对指标变量进行筛选，实现了数据降维。之后，利用灰色预测对股票价格影响指标 进行预测，并在此基础上利用神经网络模型对股票收盘价进行预测。结果表明，利用灰色系统 和BP神经网络结合的模型所得预测结果平均相对误差为0.095，且运行效率较高，对股票预测 具有一定的积极意义。

Stock Price Prediction of GEM Based on Grey System and Neural Network

Stock price forecasting has always been a hot topic in investment field. Since the stock prices is affected by many non-linear factors, it is difficult to obtain accurate prediction results. In order to eliminate the multi- collinearity of stock index, the adaptive-Lasso algorithm is used to filter the index variables, and the dimensionality of data is reduced. After that, the influence index of stock price is forecasted by grey forecast, and on this basis, the closing price of stock is forecasted by neural network model. The result shows that the average relative error of the forecasting result obtained by the combination of the gray system and the BP neural network is 0.095, and the operation efficiency is high, which has certain positive significance for stock forecasting.