利用BP神经网络系统对股票市场进行预测与分析的研究

发布时间：2018-01-03 00:08

本文关键词：利用BP神经网络系统对股票市场进行预测与分析的研究　出处：《天津大学》2013年硕士论文　论文类型：学位论文

【摘要】：股票市场是一个充满了机遇与陷阱的地方。自从1990年起，股票公开在上海、深圳两地发行以来，炒股票已经成为国人日常经济行为的中的一部分。虽然股票的收益可以非常高，但是股票同样具有着高风险。在这种形势下，越来越的投资人和投资机构开始关注于对于股市行情的走向，试图通过股票背后大量的数据来实现对股票走势的预测。在这种形势下，对股票市场内在规律的研究和预测具有着极其重要的理论意义和实用价值。本文试图通过利用BP（Back Propagation）神经网络进行对股票的分析和预测。股市是一个非常复杂的非线性的动力学系统，而神经网络具有很强的非线性逼近能力和自学、自适应等特性。通过整理股票价格历史数据，并使BP神经网络训练学习历史数据，可以有效的找到股票市场价格变动的规律，，来达到预测股票未来价格趋势的目的。本文分析了使用BP网络对股市价格进行预测分析中的原理，建立三层前馈神经网络建立对股票的预测模型。在实验中，通过对BP网络参数的调整，以达到比较好的学习效果。再以五只纳斯达克股票为例，应用已经实现的预测模型对其股价的未来走势进行预测，取得了比较好的效果。通过过往的理论研究和BP神经网络的特点，可以证明可以使用BP神经网络对股票价格进行有效预测。
[Abstract]:The stock market is a place full of opportunities and pitfalls. Since 1990, stocks have been publicly issued in Shanghai and Shenzhen. Stock speculation has become a part of people's daily economic behavior. Although the return of stocks can be very high, but stocks also have high risk. In this situation. More and more investors and investment institutions are focusing on the direction of the stock market, trying to achieve the prediction of the stock market through a lot of data behind the stock. In this situation. The research and prediction of the inherent law of stock market is of great theoretical significance and practical value. This paper attempts to analyze and predict stocks by using BP(Back Propagation neural network. Stock market is a very complex nonlinear dynamic system. The neural network has strong ability of nonlinear approximation, self-learning, self-adaptation and so on. Through sorting out historical data of stock price, BP neural network is trained to learn historical data. We can find the law of stock market price change effectively to predict the stock price trend in the future. In this paper, the principle of using BP neural network to predict the stock price is analyzed, and the three-layer feedforward neural network is established to build the forecasting model of stock. In the experiment, the parameters of BP network are adjusted. In order to achieve a better learning effect. Then take five NASDAQ stocks as an example, using the realized prediction model to predict the future trend of its stock price. Through the past theoretical research and the characteristics of BP neural network, it can be proved that the BP neural network can be used to predict the stock price effectively.
【学位授予单位】：天津大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP183;F832.51

【参考文献】