基于主成分分析与神经网络组合的股价变化趋势预测
|
摘要
股票价格影响因素众多,且存在复杂的相关关系,传统股票价格预测模型和神经网络模型无法消除影响因素直接的相关性,从而导致预测精度较低。为提高股票价格预测精度,提出了一种主成分分析与BP神经网络组合模型。利用主成分分析降低股票价格预测影响因素之间存在的冗余信息,降低BP构,提高组合模型训练速度和预测精度。对自科学研究和技术服务业相关上市企业周股价进行仿真实验,结果表明PCABP组合模型的训练速度快,预测精度高,能够为股票价格预测提供有价值的参考。
There are many influencing factors of stock price,and there are complex correlations,and the traditional stock price prediction model and neural network model can not eliminate the direct correlation of influencing factors,which leads to low prediction accuracy.In order to improve the accuracy of stock price prediction,a combination model of principal component analysis and BP neural network is proposed.The principal component analysis is used to reduce the redundant information existing between the influencing factors of stock price prediction,to reduce the dimension of BP neural network model input data,to simplify the topology of neural network,and to improve the training speed and prediction accuracy of combinatorial model.The simulation experiment of weekly stock price of listed enterprises related to scientific research and technology Service industry shows that the training speed of PCA-BP combination model is fast and the prediction precision is high,which can provide valuable reference for stock price forecast.
There are many influencing factors of stock price,and there are complex correlations,and the traditional stock price prediction model and neural network model can not eliminate the direct correlation of influencing factors,which leads to low prediction accuracy.In order to improve the accuracy of stock price prediction,a combination model of principal component analysis and BP neural network is proposed.The principal component analysis is used to reduce the redundant information existing between the influencing factors of stock price prediction,to reduce the dimension of BP neural network model input data,to simplify the topology of neural network,and to improve the training speed and prediction accuracy of combinatorial model.The simulation experiment of weekly stock price of listed enterprises related to scientific research and technology Service industry shows that the training speed of PCA-BP combination model is fast and the prediction precision is high,which can provide valuable reference for stock price forecast.
引文
[1]刘海啸,孟珊珊,杜佳豪.基于数据挖掘的股市波动与宏观经济指标的关联性研究[J].燕山大学学报,2018,19(5):89-96
[2]Mendelsohn Stein.Fundamental analysis meets the neural etwork[J].Futures,1991,20(10):22-24.
[3]霍建军.基于灰色系统和神经网络的股价指数预测:[D].哈尔滨工程大学,2006.
[4]张秀艳,徐立本.基于神经网络集成系统的股市预测模型[J]系统工程理论与实践,2003,9:67-70.
[5]孙磊平.数据挖掘方法在股票分析中的应用与研究[D].西南财经大学,2013.
[6]王刚,许晓兵.基于小波分析与神经网络时间序列的股票预测方法[J].金融经济,2013(24).
[7]张翱翔.基于BP神经网络股指预测系统的国内外股市预测研究[D].湖南大学,2017.
[8]孙彬,李铁克,王柏琳.基于股票市场灵敏度分析的神经网络预测模型[J].计算机工程与应用,2011,47(1):26-31.
[9]兰强太.基于主成分分析与BP神经网络算法[D].暨南大学,2017.
[10]江建明,柳长青.基于主成分分析法与BP神经网络的股指期货价格预测[J].金融视线,2015,18.093:176-177.
[11]韩力群.人工神经网络教程[M].北京邮电大学出版社,2006.
[2]Mendelsohn Stein.Fundamental analysis meets the neural etwork[J].Futures,1991,20(10):22-24.
[3]霍建军.基于灰色系统和神经网络的股价指数预测:[D].哈尔滨工程大学,2006.
[4]张秀艳,徐立本.基于神经网络集成系统的股市预测模型[J]系统工程理论与实践,2003,9:67-70.
[5]孙磊平.数据挖掘方法在股票分析中的应用与研究[D].西南财经大学,2013.
[6]王刚,许晓兵.基于小波分析与神经网络时间序列的股票预测方法[J].金融经济,2013(24).
[7]张翱翔.基于BP神经网络股指预测系统的国内外股市预测研究[D].湖南大学,2017.
[8]孙彬,李铁克,王柏琳.基于股票市场灵敏度分析的神经网络预测模型[J].计算机工程与应用,2011,47(1):26-31.
[9]兰强太.基于主成分分析与BP神经网络算法[D].暨南大学,2017.
[10]江建明,柳长青.基于主成分分析法与BP神经网络的股指期货价格预测[J].金融视线,2015,18.093:176-177.
[11]韩力群.人工神经网络教程[M].北京邮电大学出版社,2006.