基于改进的灰色Markov链的预测方法及其实证

文章将灰色预测模型GM(1,1)与Markov链方法相结合,建立灰色Markov预测模型,预测波动上升的数据系列,改进了状态转移矩阵的传统计算方法,对状态施行动态滑动转移,用更新的一步状态转移概率矩阵,对未来数据所处的状态进行一步或多步预测.在实证中,以黄石市农民年人均纯收入为例,把农民收入看作一个灰色系统,农民人均收入看作灰色量,通过过去年份的数据挖掘,在Mathematica7.0环境下,给出了2009年黄石农民年人均纯收入的状态预测,并提出影响预测精度的若干因素.

Prediction Method and Application Based on Improved Gray Markov Chain

This paper integrates the gray prediction model GM （1,1） with Markov chain approach and constructs the gray Markov prediction model to predict the rising volatility data series, which has improved the traditional calculation method of the state transition matrix. The predition method implements dynamic sliding transfer to the state and by u- sing the updated step of the transition probability matrix for the state, it predicts the one step or multi - step of the future data state. Taking the per capita income of the farmers in Huangshi City for example, the income of the farmers as a gray system, and the per capita income as the amount of gray, through the last several years data mining, in the Mathematica7. 0, the state prediction of the per capita net income of Huangshi farmers in 2009 is given and a series of factors affecting the prediction accuracy are put forward.