BP神经网络在证券指数预测中的研究与应用

发布时间：2018-05-05 03:38

  本文选题：股票指数预测 + 神经网络　； 参考：《大连海事大学》2012年硕士论文

【摘要】：证券是一个高度复杂的非线性动态系统,其变化规律既有一定的自身的趋势性,又受政治、经济、心理等诸多因素的影响。证券市场的高风险高收益一直倍受投资者的关注,为了获取丰厚的收益,如何建立一个成功率比较高的预测理论和模型,成为了学术界研究的热点问题。 已有的研究工作是从技术分析、心理分析以及数理统计为基础的的时间序列定量预测方法,然而其在股市的研究中面临着许多困难,难以取得满意的效果。随着证券市场混沌和分形理论的逐步确立,人们开始利用神经网络对证券市场的变动加以预测。神经网络具有自组织、自适应和学习非线性映射强等特点,能自动从历史数据中提取有关经济活动中的知识,非常适合应用于经济领域的信息处理以及分析时间序列,对于解决股票预测领域中的一些问题比较实用。BP(Back Propagation)网络是一种被广泛运用的神经网络,它的核心是BP算法,是一种对于众多基本子系统构成的大系统进行计算的严格而有效的方法。 本文首先分析了股票预测的需求,并介绍了神经网络的基本原理和方法。结合股票市场中的实际问题,提出了基于BP人工神经网络的股票预测模型,实现了三层的BP网络来对上海交易所上证指数进行预测。 最后通过实验分析证明了BP神经网络不仅能够学习训练集的例子,且能从训练集中提炼出某种一般性原理、规律,具有较强的非线性函数拟合特性,因此对于预测短周期内股指波动有较强的适用性。
[Abstract]:Securities is a highly complex nonlinear dynamic system, its change law has its own tendency, and is influenced by many factors such as politics, economy, psychology and so on. The high risk and high yield of the securities market has been paid much attention by investors. In order to obtain rich income, how to establish a prediction theory and model with high success rate has become a hot issue in academic circles. The existing research work is based on technical analysis, psychological analysis and mathematical statistics. However, it faces many difficulties in the research of stock market, and it is difficult to obtain satisfactory results. With the gradual establishment of chaos and fractal theory in securities market, people begin to use neural network to predict the change of securities market. Neural network has the characteristics of self-organization, self-adaptation and strong learning nonlinear mapping. It can automatically extract knowledge about economic activities from historical data. It is very suitable for information processing and analysis of time series in the economic field. For solving some problems in the field of stock forecasting, the BP back propagation network is a widely used neural network, the core of which is BP algorithm. It is a strict and effective method for computing large scale systems composed of many basic subsystems. This paper first analyzes the demand of stock forecasting, and introduces the basic principle and method of neural network. Combined with the practical problems in the stock market, a stock forecasting model based on BP artificial neural network is proposed, and a three-layer BP network is implemented to predict the Shanghai Stock Exchange index. Finally, it is proved that BP neural network can not only learn the example of training set, but also extract some general principles and rules from the training set, and has strong nonlinear function fitting characteristic. Therefore, it has strong applicability to predict the volatility of stock index in short period.
【学位授予单位】：大连海事大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F830.91;TP183

【相似文献】

中国期刊全文数据库 前10条

1 戚杰,周敬宣,陈云峰,李艳萍;区域环境-经济系统的动态仿真——以南水北调后的湖北省襄樊市为例[J];系统工程理论与实践;2004年03期

2 杨楠;趋势外推法在房价指数预测中的应用[J];统计与预测;2004年01期

3 张神勇;从商品零售物价指数预测看保值补贴率的变化趋势[J];上海统计;1996年04期

4 徐静芳;1999年上海物价指数预测与调控[J];上海综合经济;1999年01期

5 王悦;人工神经网络在经济预测中的应用[J];北京广播电视大学学报;2002年01期

6 王渡;;基于改进的BP网络的循环流化床锅炉建模[J];科技创新导报;2009年12期

7 刘丹;;基于一种改进的BP神经网络的飞机目标识别[J];中国高新技术企业;2011年16期

8 冯居易;;神经网络在股票价格预测中的建模及应用[J];福建电脑;2009年01期

9 贺艳辉;袁永明;张红燕;龚，

本文编号：1846048