基于三次样条权函数神经网络的股价预测  

Stock Prediction Based on Neural Networks with Cubic Spline Weight Functions

作　　者：张浩[1] 张代远[1,2,3] 

机构地区：[1]南京邮电大学计算机学院,江苏南京210003 [2]江苏省无线传感网高技术研究重点实验室,江苏南京210003 [3]南京邮电大学计算机技术研究所,江苏南京210003

出　　处：《计算机技术与发展》2014年第6期28-31,共4页Computer Technology and Development

基　　金：江苏高校优势学科建设工程资助项目(yx002001)

摘　　要：随着经济的发展,股票投资已成为很多人的一种投资理财方式,而股票价格的预测也成为投资者关心和研究的焦点。建立一个运算速度和精确度都比较高的股价预测模型,对于金融投资者具有理论指导意义和实际应用价值。文中针对传统BP算法存在的学习速度慢、容易陷入局部极小值、隐层数不易确定等问题,使用三次样条权函数神经网络建立股价预测模型,克服了传统神经网络的缺点。仿真结果表明,该模型具有较高的预测精度,能够对股市进行有效的预测。With economic development,stocks have become a way of finance and investment for many people,and the stock price forecasting has become the focus of investors＇ attention and study.Establishing a stock price forecasting model with high computing speed and accuracy has theoretical and practical significance for financial investors.As the traditional BP algorithm has problems such as low learning speed,easy to fall into local minimum value,difficult to determine the number of hidden layer neurons,the stock prediction model is established using the neural networks with cubic spline weight functions to overcome the shortcomings of traditional neural networks.The simulation results show that the model has high accuracy and can effectively predict the stock market.

关 键 词：权函数 三次样条函数 BP算法 神经网络 股价预测 

分 类 号：TP31[自动化与计算机技术—计算机软件与理论]