灰色模型及其组合模型在证券分析的应用

【作者】 李攀峰；

【导师】 唐应辉；

【作者基本信息】 电子科技大学 ， 运筹学与控制论， 2002， 硕士

【摘要】 本文主要讨论了灰色模型应用于证券分析的可行性，并通过对灰色模型的改进，建立了组合模型，以及新的分析方法：均线形态组合法，在一定程度上实现了对股票价格的中、短期趋势都有较好的预测效果。 第二章对灰色模型应用于证券分析的可行性进行论证，改进了残差灰色预测模型，并通过对灰色系统建模数据的变换，改进了数据的光滑特性，进一步提高预测的精度。 第三章将人工神经网络引入证券分析中，通过选取适当的网络结构、训练参数和计算机编程实现了较好预测效果。 第四章应用结构模式识别的方法，建立多阶均线形态组合序列。通过对股市历史数据的变换和分析，考查各形态组合序列对应的盈利和亏损概率，建立新的证券投资分析方法即均线形态组合预测法。其优点是可以部分弥补传统技术分析中的滞后性和不确定性，能根据市场的变化及时调整相应的参数指标，有效提高预测的成功率，并易用计算机作出识别。 第五章针对证券数据序列同时具有增长性和波动性的特性，建立组合灰色神经网络模型，研究了同时考虑两种非线性趋势的预测问题。 本文通过大量的实例，对以上方法进行了验证。更多还原

【Abstract】 In this paper, we investigate the feasibility of Grey Model applied to security analysis, improve Grey Model, build compositional model and new security analysis method, namely equal-string form combination predicted method. In some extent, the preferable predictive validity of the long-term and short-term current of securities price is arrived.In Chapter 2, we improve the residual forecast Grey Model, by transformation of the primary data, the lubricity of data is improved and the forecast precision is raised.In Chapter 3, we introduce Artificial Neural Network (ANN) into security analysis, select the well-founded net structure and training parameter, andprogram. The predictive validity is good.In Chapter 4, we use the method of structural pattern recognition, build multistage sequence of equal-string form. By transforming and analyzing history data of stock market, examine the probability of payoff or loss. Based on that, a new analysis method of securities investment is established, namely Equal-String Form Combination predicted method. Its advantage is remedy lag and uncertainty of traditional analysis, adjust relevant parameter to consistent with the movement of market immediately, heighten success ratio of forecast, and recognized easily by computer.In Chapter 5, Because the data sequence has double characteristic of increasing and fluctuating, we put forward the combined grey neural network model, study the forecast problem of double nonlinear trend.Using several examples, we validate these method.更多还原