基于支持向量机的股价预测方法及实证研究-概率论与数理统计专业论文.docxVIP

摘要随着中国经济的飞速发展，中国的股票市场越来越趋于完善， 摘要 随着中国经济的飞速发展，中国的股票市场越来越趋于完善， 股票交易已经成为人们理财的重要手段，对股票价格的预测也越来 越引发人们的关注，所以如何准确的预测股票价格以及波动趋势成 为了一个重要的研究方向，国内外很多学者都致力于此方面的研究。 支持向量机是机器学习的一种新理论，是基于统计学习理论并 利用最优化方法来解决学习问题的新方法。本文在分析股票市场特 点的基础上，介绍了常用的时间序列预测方法并提出利用支持向量 机建立回归预测模型对股价进行预测。我们选取了长安汽车2007年 1月1010年1月的开盘价作为研究对象，利用3种不同的参数选 择方法建立了支持向量机模型并进行了实证检验。 此外，为了验证支持向量机模型的预测效果，我们利用ARIMA、 灰色理论以及神经网络模型对长安汽车的开盘价进行预测，并与支 持向量机模型的预测结果做了对比分析。 最后，为了更好的预测股票价格的波动趋势，我们将信息粒化 方法与支持向量机模型结合起来，对股票价格的变化范围进行了预 测。结果表明，支持向量机模型不仅对股票价格的预测有着良好的 准确性，对股票价格波动趋势的预测也有着很好的效果。 关键词股票价格，支持向量机，ARIMA，核函数，参数选择 ABSTRACT ABSTRACT With the development at full speed of China’S economy,the more perfect the stock market of China becomes more and more，the stock exchange has already become the important means of people’S financing， cause people’S conccm to the prediction of the stock price more and more too，SO how accurate prediction stock price and trend of fluctuating become all important research direction，a lot of domestic and international scholars are devoted to the research of this respect． It is a kind of new theory of machine study to support vector machine，is on the basis of counting the theory of studying and utilizing optimizing the method and solving the new method of the problem concerning study most．This text is on the basis of analysing characteristic of stock market，have recommended the commonly used time array to predict the method and propose utilizing the vector quantity supporting machine and set up predicting that models predict the stock price．We have chosen the opening price of January of 2007．January of 20 1 0 of automobile of Chang An as the research object，utilize 3 kinds of di fferent parameters to choose the method to set up the model of vector quantity supporting machine and examine the real example． In addition，in order to prove that the prediction result of supports vector machine model，we utilize ARIMA，grey theory and neural Il network network model to predict the opening price of the automobile of Chang