中文摘要：

为了解决目前常用预测模型对随机波动性较大数据预测精度偏低的问题,文章在灰色预测GM（1,1）的基础上引入马尔可夫状态转移矩阵,建立了灰色马尔可夫预测模型（GMM）,并将该法运用到煤矿顶板致死人数的预测中。经计算GMM模型的预测平均相对误差为1.181%,最大相对误差3.426%,与GM（1,1）法相比,后者精度分别提高了21倍和13倍。

英文摘要：

In order to solve the problem that the forecast accuracy of commen prediction model is low for the data which has a large random volatility,the Markov state shift matrix was introduced into the Grey prediction,and the Grey Markov prediction model（GMM） was established. The method is applied in the prediction of the death from the roof accident,the average relative error of the prediction is 1. 181% and the maximum relative error is 3. 426%,which has increased by 21 times and 13 times respectively compared with GM（1,1） method.