文件名称:vol
介绍说明－－下载内容均来自于网络，请自行研究使用
matlab金融时间序列ARMA建模
结果分析：
1.预测结果从第四步开始，预测值不再改变，因为ARMA是收敛的回归模型，而我们做的工作并不是模拟，所以，当预测步长足够长时，它最终将收敛于一个不变得预测值
2.既然预测值一样，为什么还原为成交量后，在置信区间下预测的最大值与预测均值的差比预测均值与最小值的差要大？因为将对数差分值还原时，需用到的指数函数为凹函数-matlab Financial Time Series the the ARMA modeling results Analysis: 1. predicted results from the fourth step, the predicted values no longer change, because the the ARMA convergence regression model, and the work we do is not analog, so when the prediction step enough after a long time, it will eventually converge to a become predictive value since the predicted value, why restore volume, maximum predicted mean forecast in the confidence interval for the difference than forecast average with the minimum difference big? Because the logarithmic differential value reduction required to exponential function is a concave function
结果分析：
1.预测结果从第四步开始，预测值不再改变，因为ARMA是收敛的回归模型，而我们做的工作并不是模拟，所以，当预测步长足够长时，它最终将收敛于一个不变得预测值
2.既然预测值一样，为什么还原为成交量后，在置信区间下预测的最大值与预测均值的差比预测均值与最小值的差要大？因为将对数差分值还原时，需用到的指数函数为凹函数-matlab Financial Time Series the the ARMA modeling results Analysis: 1. predicted results from the fourth step, the predicted values no longer change, because the the ARMA convergence regression model, and the work we do is not analog, so when the prediction step enough after a long time, it will eventually converge to a become predictive value since the predicted value, why restore volume, maximum predicted mean forecast in the confidence interval for the difference than forecast average with the minimum difference big? Because the logarithmic differential value reduction required to exponential function is a concave function
(系统自动生成,下载前可以参看下载内容)
下载文件列表
vol.m
SH000001.TXT