股票分类指数的多元马尔可夫链模型
发布时间：2018-11-24 07:33
【摘要】：针对多元马尔可夫链模型分析中,待估参数数量大而导致估计困难的问题,文章提出了改进方法。以一元预测误差最小为优化目标,对列间权重参数进行了分批次优化求解。应用多元马尔可夫链模型,对股票五大行业分类指数序列进行建模,研究了各行业分类指数相互间的内在相依特征。应用数据序列间的多元交互信息,对行业分类股指序列进行了预测。
[Abstract]:In order to solve the problem of large number of parameters to be estimated in multivariate Markov chain model analysis, an improved method is proposed in this paper. With the minimum prediction error as the optimization objective, the weight parameters between columns are solved by batch optimization. In this paper, the multiple Markov chain model is used to model the stock classification index series of five major industries, and the internal dependence characteristics of each industry classification index are studied. This paper predicts the industry classification stock index series by using the multiple interactive information between the data series.
【作者单位】： 重庆大学经济与工商管理学院;重庆大学数学与统计学院;
【基金】：国家自然科学基金资助项目(71171209)
【分类号】：F830.91;F224
本文编号：2352802
[Abstract]:In order to solve the problem of large number of parameters to be estimated in multivariate Markov chain model analysis, an improved method is proposed in this paper. With the minimum prediction error as the optimization objective, the weight parameters between columns are solved by batch optimization. In this paper, the multiple Markov chain model is used to model the stock classification index series of five major industries, and the internal dependence characteristics of each industry classification index are studied. This paper predicts the industry classification stock index series by using the multiple interactive information between the data series.
【作者单位】： 重庆大学经济与工商管理学院;重庆大学数学与统计学院;
【基金】：国家自然科学基金资助项目(71171209)
【分类号】：F830.91;F224
【二级参考文献】
相关期刊论文 前4条
1 侯木舟,韩旭里;基于MATLAB的神经网络在股市预测中的应用[J];系统工程;2003年02期
2 李一军,叶强;基于BP神经网络的产品报价决策支持方法及其报价风险研究[J];管理工程学报;2003年04期
3 禹建丽,孙增圻,Valeri.Kroumov,成久洋之,刘治军;基于BP神经网络的股市建模与决策[J];系统工程理论与实践;2003年05期
4 张秀艳,徐立本;基于神经网络集成系统的股市预测模型[J];系统工程理论与实践;2003年09期
，本文编号：2352802
本文链接：https://www.wllwen.com/jingjilunwen/guojijinrong/2352802.html
