基于支持向量机的对优质股票选取的研究

发布时间：2018-11-07 11:48

【摘要】：伴随经济的发展和人们收入的提高，股票市场受到越来越多的人的关注，股票投资己经成为现代人生活中一个非常重要的组成部分。而到目前为止，随机挑选股票已经不太可能得到回报，且伴随着有效市场假说理论的推广，消极投资策略逐渐被广大投资者认可。但不可否认，在股票市场中总有一些股票价值正被低估，因此，在足够竞争性的市场上，投资者想获得的收益大于市场上的平均收益，那么对于股市预测方法的研究不但具有深刻的理论意义更具有极其重要的应用价值。同时，经过多年的发展，股票市场积累了大量的数据信息，如何将其应用于投资决策中便成为了股市预测的焦点。支持向量机是数据挖掘中的一项新技术。本文则将支持向量机分类机应用于股票的选取上。首先本文介绍了介绍了统计学习理论和建立在其上的支持向量机方法，详细描述了支持向量机方法的基本原理。其次，为了提高股票选取的精度，本文构建了财务舞弊识别的SVM模型，目的在于及早的发现上市公司财务状况的异常，对投资者进行预警，从而保护投资者的利益。而且，在此模型中，打破了传统的1：1的配对方法，，采用主成分分析方法获得控制样本，并且引入了支持向量机对非平衡数据的处理方法，得到了较高的准确率。最后，本文对深沪100只股票进行选取，结果令人满意。由于支持向量机自身所具备的优势，随着研究与实践的不断深入，它必定能够在股市预测方面取得新的突破与进展。
[Abstract]:With the development of economy and the improvement of people's income, more and more people pay attention to the stock market. Stock investment has become a very important part of modern people's life. Up to now, it is unlikely to get a return on the random stock selection, and with the promotion of efficient market hypothesis, the passive investment strategy is gradually accepted by the majority of investors. But it is undeniable that there are some stocks that are undervalued in the stock market, so that in a sufficiently competitive market, investors want to earn more than the average return on the market. So the research of stock market forecasting method not only has profound theoretical significance, but also has extremely important application value. At the same time, after years of development, the stock market has accumulated a large amount of data information, how to apply it to investment decisions has become the focus of stock market forecasting. Support vector machine (SVM) is a new technique in data mining. In this paper, support vector machine classifier is applied to stock selection. Firstly, this paper introduces the statistical learning theory and the support vector machine (SVM) method based on it, and describes the basic principle of the SVM method in detail. Secondly, in order to improve the accuracy of stock selection, this paper constructs the SVM model of financial fraud identification, the purpose of this model is to detect the abnormal financial situation of listed companies as early as possible, to warn investors, so as to protect the interests of investors. In this model, the traditional 1:1 pairing method is broken, the control samples are obtained by principal component analysis (PCA), and the support vector machine (SVM) is introduced to deal with the unbalanced data. Finally, 100 stocks in Shenzhen and Shanghai are selected and the results are satisfactory. With the development of research and practice, support vector machine (SVM) will be able to make new breakthrough and progress in stock market prediction.
【学位授予单位】：重庆交通大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F832.51;TP18

【相似文献】