毕业论文神经网络在股价预测中的应用.doc

本科毕业论文(设计) 题 目：____神经网络在股价预测中的应用 _______ Neural Networks Application in Stock Price Forecsting 学 院：__ _自动化工程学院_____________ 专 业： 自动化_______________ 姓 名： 指导教师：_______ _ ________________ 2013年 6 月 5 日 摘 要 随着全球经济的高速发展和各国人们对于股票市场投资认识的不断加深，越来越多的人们踏入了股市投资这个行业，因此更多的投资者便越来越关心股票价格预测这个话题。由于股票投资一个最大的特点是更大的风险就意味着更大的利益，因此股票投资者希望通过建立一个非常精确的股市价格的预测模型来获得更多的利益，故股市价格的预测具有非常重要的应用价值。 但是由于股票市场是一个十分繁杂的非线性的动力学系统，影响股票市场变化的因素除了其自身所包含的规律之外，还包括经济环境等诸多因素，故一般的预测方法难以对其进行准确的预测。由于BP神经网络具有非常好的非线性映射能力、泛化能力和容纳错误的能力，因而其非常适合用来解决股票预测领域中的一些问题。 在本文中，详细介绍了BP算法的基本知识，同时针对BP算法在股市预测中存在的学习速度慢、容易陷入局部极小值、预测结果精度不高等问题，提出一种改进的BP神经网络算法。然后本文基于BP网络进行股市的价格进行预测分析的原理，建立一个基于BP网络的股市价格的预测模型，然后采用改进后的BP算法进行股市预测，并通过MATLAB软件对其预测过程进行仿真实验，最终结果达到了所预料的效果。 关键词：神经网络，股票预测，BP算法 Abstract With the increasingly rapid economic development and the change of peoples investment consciousness, stock investment has become an important part of modern life, and the prediction of stock price has become the focus of concern and research of investors. How to build a higher computing speed and a more accurate prediction model of stock market, not only has theoretical significance for financial investors, but also has a more important practical application value. Because the income of stock investment is often proportional to the risk. 　　　Although the stock market is a highly complicated nonlinear system, its variation has its own regulation, it is influenced by the market, the macroeconomic environment and other noneconomic reasons, so ,it is hard to forecast the accuracy of the method. The BP neural is suitable to solve some problems in stock market prediction, because of its nonlinear mapping ability, generalization ability and good tolerance,　　　 In this thesis, the author introduces the basic knowledge of BP algorithm, as well as an improved algorithm of BP neural network which is in order to s