基于灰色系统理论的股票价格预测研究

发布时间：2019-04-08 18:01

【摘要】：本研究旨在利用灰色系统理论研究证券股票市场发展的一般规律，通过优化的灰色模型，即Grey Model模型对股票价格预测进行研究。随着证券市场的不断发展和完善，投资者对股票价格走势的认识也逐渐加深，渐渐的产生出多种股票价格预测方法。股票市场瞬息万变，传统的方法因其缺陷不能有效及准确地对股价进行预测，，因此，应该探索和研究更佳有效的预测股票价格走势的复杂性及规律性的方法。本研究首先要建立灰色系统理论的模型，接着对灰色系统的功能以及系统的各个因素间的关系等进行具体的量化研究。由于股票走势的复杂性，该理论还不能充分满足股价预测的需要。为了进一步提高灰色模型预测的精度，本文对GM（1,1）模型的残差进行修正。利用灰色GM（1,1）模型的剩余残差建立模型，并将该模型用于股价预测。本研究采取残差修正模型，以三爱富这只股票作为研究对象，首先通过GM（1,1）模型计算得出模拟序列，并计算其残差序列。由于模型要求的原始数列是非负序列，本研究者给残差序列加上其最小值的绝对值，使其成为满足条件的非负序列。然后运用改进后的残差修正模型得出修正后的预测值。最后通过多种检验方法验证了模型的合格性。研究显示：修正前的GM（1,1）模型预测的平均相对误差为0.01272；而改进后的GM（1,1）预测的平均相对误差为0.01146。实验证明，残差修正模型的预测效果要比不带残差修正的GM模型预测的效果更好。然而，股票买卖操作具有其复杂多样性，如何更加有效的对GM模型进行优化，使其为股票投资者的买卖操作提供更精确的预测，仍将是今后的一个重要课题。
[Abstract]:The purpose of this study is to use the grey system theory to study the general law of the development of the stock market, and to study the stock price prediction through the optimized grey model, that is, the Grey Model model. With the continuous development and perfection of the stock market, investors have a deeper understanding of the stock price trend, and gradually produce a variety of stock price forecasting methods. Because the stock market changes rapidly, the traditional method can not predict the stock price effectively and accurately because of its defect. Therefore, it is necessary to explore and study more effective methods to predict the complexity and regularity of stock price trend. Firstly, the model of grey system theory should be established, and then the function of grey system and the relationship among the various factors of grey system should be studied in detail. Because of the complexity of stock trend, the theory can not fully meet the demand of stock price forecast. In order to improve the precision of grey model prediction, the residual error of GM (1,1) model is corrected in this paper. The residual of the grey GM (1,1) model is used to build the model, and the model is used to forecast the stock price. In this study, the residual correction model was adopted and the stock of Sanai Fu was taken as the research object. Firstly, the simulated sequence was calculated by GM (1,1) model, and the residual sequence was calculated. Since the original sequence required by the model is a non-negative sequence, the author adds the absolute value of its minimum value to the residual sequence to make it a non-negative sequence which satisfies the conditions. Then the modified residual correction model is used to obtain the modified predictive value. Finally, the qualification of the model is verified by a variety of testing methods. The results show that the average relative error of the modified GM (1,1) model is 0.01272, and the average relative error of the improved GM (1,1) prediction is 0.01146. Experiments show that the prediction effect of residual correction model is better than that of GM model without residual correction. However, stock trading operations have its complexity and diversity. How to optimize the GM model more effectively so as to provide more accurate prediction for the trading operations of stock investors will still be an important topic in the future.
【学位授予单位】：重庆师范大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F830.91

【相似文献】