论文部分内容阅读
将黄金数据的尖峰厚尾、异方差性及杠杆效应等统计特征与马尔科夫概率转移矩阵所具有的动态变化规律结合,提出一种改进的灰色马尔科夫模型.模型首先对数据进行统计分析,建立相应的概率统计模型并用此模型对系统发展变化趋势进行拟合.在拟合序列的基础上利用马尔科夫链的动态转移变化建立状态转移概率矩阵,采用动态数据驱动原理对未来每一步数据进行动态预测.模型既是统计方法与数据动态驱动的结合,克服了传统的灰色马尔科夫模型中对数据内在统计规律的忽视,实证表明其预测精度较灰色马尔科夫模型预测高,具有较好的实用性.
Combining the statistical characteristics of the peak and tail, heteroscedasticity and leverage of gold data with the dynamic change rule of Markov probability transfer matrix, an improved gray Markov model is proposed.At first, the data are statistically analyzed , And set up the corresponding probabilistic statistical model and use this model to fit the trend of system development.The state transition probability matrix is established by using the dynamic transition of Markov chain based on the fitted sequence and the dynamic data driven principle is applied to each step in the future Data are dynamically predicted.The model not only ignores the statistical rules of data in the traditional Gray Markov model but also proves that the prediction accuracy is higher than that predicted by the gray Markov model, Good practicality.