基于多技术指标模型的沪深300指数走势预测

发布时间：2018-01-08 19:00

本文关键词：基于多技术指标模型的沪深300指数走势预测　出处：《江西财经大学》2012年硕士论文　论文类型：学位论文

【摘要】：沪深300指数反映了中国证券市场股票价格变动的概貌和运行状况,能够作为投资业绩的评价标准,越来越受到投资者的青睐。在技术分析中具有重要代表性的技术指标能不能对股市进行预测,存在着股市是否达到弱式有效性的问题。本文综述了研究我国股市有效性问题的相关文献,利用ADF单位根检验得出我国股市未达到弱式有效性,在这一前提下,基于多技术指标构建模型短期预测沪深300指数。股票中的技术指标,是评价股票某一特性而构造出的数学公式,用来计算股票相关数据。技术指标分析法,根据统计学中分析方法,考察技术指标间的统计性质,构建模型预测股票未来走势的分析方法。本文根据技术指标选取的综合性与系统性原则、科学性原则、可操作性原则和组合使用原则,挑选出能概括超买超卖型、成交量型、能量型、趋势型和停损型的14个常用技术指标,对14个技术指标提示的买卖点进行数据预处理,以便后文分析。由于单一技术指标在提示股票指数买卖点上存在着片面性,利用多个技术指标提高预测准确率就显得势在必行,而且选取技术指标的方式对于成功构建预测模型预测沪深300指数走势至关重要。本文运用统计学中的泊松相关系数矩阵考察技术指标两两之间在提示买卖点上的相似性,再进一步利用聚类分析对技术指标提示买卖点进行分类,最后利用灰色关联度分析对常用的14个技术指标与沪深300指数之间的关联程度数量化,进行排名并结合样本数据的实际状况,最终选取与沪深300指数关联度比较大的OBV、RSI、PSY、DMI、SAR五个技术指标,基于这五个技术指标构建预测模型短期预测沪深300指数,在最大精准率下应用最少的技术指标是本文选取技术指标的一个基本原则。预测方法按统计性质可分为定性预测和定量预测,本文主要是对沪深300指数运用定量分析手段预测其短期走势。定量预测方法的发展根据出现时间的先后大体上可分为三个阶段：结构计量模型阶段、时间序列分析阶段和智能预测阶段。由于股票市场无时无刻都受到各种确定或不确定性因素的影响,并且时间的不可逆性导致了股票市场具有非线性的特征,继续使用以前的线性分析或近似分析已无法准确分析研究出股票市场的特征和趋势。从定量预测发展阶段来说,目前主要研究集中在非线性、非参数的智能预测,把新的预测方法应用于实际是否能提高预测效果和精度就显得异常重要。股市是一个复杂的非线性动态系统,具有非线性和时变性等特征,本文在对股价主要预测方法介绍及评论后,最终确定决策树分析和RBF神经网络分析预测沪深300指数。决策树分析不仅能对沪深300指数走势方向进行预测,而且能够验证技术指标用于预测的有效性。最后运用RBF神经网络分析对沪深300指数短期具体点位进行预测。实证分析表明决策树分析和RBF网络分析能够准确地进行短期预测,为投资者短期预测提供思路及方法参考。最后本文分别给予证券监管机构和投资者相关建议,对证券监管机构来说提高我国证券市场的有效性,关键在于建立信息披露制度,保护投资者利益,促进上市公司的资源优化配置。对投资者来说,要将本文分析的思路、方法和结果应用于实际操作中,投资者应关注以下几个方面：(1)基本分析与技术分析结合运用；(2)顺势而为；(3)量价配合；(4)多种技术指标结合使用；(5)利用非线性方法预测。为保证我国股市能够持续稳定的发展,不断提高股市有效性、甄别筛选技术指标思路和应用恰当的预测方法提高预测精度,具有较强的现实意义和一定的实用价值。
[Abstract]:Shanghai and Shenzhen 300 index reflects the Chinese stock market stock price situation and operating conditions, can be used as the evaluation standard for investment performance, more and more investors. Technical indicators representative in technical analysis can predict the stock market, the stock market is to reach weak efficiency. This paper a review of the relevant literature on the effectiveness of China's stock market, using the ADF unit root test that China's stock market has not reached the weak efficiency, in this premise, the construction of Shanghai Shenzhen 300 index forecast model based on multi technology index.
Technical indexes in stock, construct a mathematical formula for the evaluation of certain features of the stock, used to calculate stock data. Technical index analysis method, based on the analysis of statistical methods, statistical analysis on nature of technical indicators. The analysis method to construct a model to forecast future stock trend. Based on the comprehensive and systematic principle and technology index selection, scientific principles, operational principles and the combination principle, the selection can be summarized OBOS type, volume type, power type, 14 commonly used technical indexes and stop the trend, to suggest that the 14 technical indexes of the trading point for data preprocessing for later analysis.
Due to the single technical indicators to forecast the stock index trading point on one sidedness, the use of a number of technical indicators improve the accuracy of prediction is imperative, and the selection of technical indicators for the success of the model to predict the trend of the CSI 300 index is very important. This paper use Poisson correlation coefficient matrix in statistics on technical indicators 22 between the tips of similarity trading points, then use clustering analysis to classify technical indicators suggested that the point of sale, finally using grey relational analysis to analyze the degree of correlation between the number of 14 technical indicators and the Shanghai and Shenzhen 300 commonly used index, ranking and combined with the actual situation of the sample data, and finally selected the Shanghai and Shenzhen 300 index correlation of high OBV, RSI, PSY, DMI, SAR five technical indicators, build prediction model for short-term forecasting of Shanghai and Shenzhen 300 index based on the five technology The application of the least technical index to the maximum precision is a basic principle of selecting technical indicators in this paper.
According to the statistical properties of prediction methods can be divided into qualitative forecast and quantitative forecast, this paper is mainly on the Shanghai and Shenzhen 300 index by means of quantitative analysis to predict the short-term trend. Quantitative prediction method of development according to the times can be divided into three stages: the stage of structural econometric model, time series analysis stage and intelligent prediction stage. Because the stock market is influenced by various effects or determined every hour and moment of uncertainty, and the irreversibility of time led to the stock market has nonlinear characteristics, continue to use the previous linear analysis or similar analysis has been unable to accurately analyze the characteristics and trends of the stock market. From the quantitative prediction of the development stage, the main research focus on nonlinear, non parametric intelligent prediction, the application of the new forecasting method to the actual effect and can improve the prediction accuracy is abnormal It's important.
The stock market is a complex nonlinear dynamic system with nonlinear and time-varying characteristics, based on the introduction and comment on the main stock price prediction method, final decision tree analysis and RBF neural network prediction and analysis of the CSI 300 index. The decision tree analysis can not only predict the trend of the CSI 300 index, effective and can be used to prediction verification technology index. Finally using RBF neural network analysis on the Shanghai and Shenzhen 300 index short-term specific point prediction. The empirical analysis shows that the decision tree analysis and RBF network analysis can accurately predict and provide reference ideas and methods to forecast short-term investors.
At the end of this paper were given the securities regulators and investors related suggestions, improve the effectiveness of China's securities market of securities regulators, the key lies in the establishment of information disclosure system, to protect the interests of investors, promoting the optimal allocation of resources of listed companies. For investors, to the analysis of the ideas, methods and results are applied to the actual operation, investors should focus on the following aspects: (1) the fundamental analysis and technical analysis combined with the use of the flow; (2); (3) with volume price; (4) use a variety of technical indicators; (5) using nonlinear prediction method. For the development of China's stock market to ensure sustained and stable, continuously improve the stock market efficiency to improve the prediction accuracy of screening, screening technology index methods and application of appropriate forecasting methods, it has a strong practical significance and practical value.

【学位授予单位】：江西财经大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F832.51

【参考文献】