组合模型在股票价格预测中应用研究  

Research on Stock Price Prediction Based on Combined Model

作　　者：王晴[1] 

机构地区：[1]南通航运职业技术学院管理信息系,江苏南通226010

出　　处：《计算机仿真》2010年第12期361-364,共4页Computer Simulation

摘　　要：研究股票价格预测问题,针对股票价格的影响因子之间存在高度冗余、非线性特点,从而导致传统方法预测准确率低的难题,为提高预测精度,提出一种支持向量机(SVM)和自回归(CAR)组合的股票价格预测方法(CAR-SVM)。CAR-SVM首先利用SVM对股票价格数据时滞阶数进行确定,然后用SVM对股票价格变化影响因素进行筛选,建立通过交叉验证得到股票价格的CAR-SVM预测模型,并以中兴通讯股票历史数据对CAR-SVM模型进行了验证性实验。实验结果表明,相对于传统的股票预测方法,CAR-SVM模型的预测精度更高,并更能反映股票价格变化的非线性动态规律,在股票价格预测中有着广泛的应用前景。The paper deals with stock price prediction.Aiming at the problems that stock price impact factors exist among high redundancy,nonlinear characteristics,leading to the traditional methods,the problem of low accuracy forecast is proposed based on support vector machine（SVM） autoregressive（CAR） stock time prediction method（CAR-SVM）.CAR-SVM uses SVM model to decide the stock time sequence order,then uses the SVM to decide the influence factors of time sequence,and gets stock time sequence of SVM forecasting model through screening cross-validation.Stock history data SVM model is verified by the experiments.Experimental results show that,compared with traditional stock forecast method,the forecasting precision of CAR-SVM model of is high,and the stock price changes to reflect the nonlinear dynamic rule.It has wide application prospects in stock price prediction.

关 键 词：支持向量机 股票价格 预测 参数选择 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]