天津市地面沉降的灰色系统—马尔柯夫预测模型应用

地面沉降速率监测值多呈波形动态，数据波动给预测带来了困难。本文以天津地面沉降多个水准点的监测数据为基本序列，介绍了传统灰色系统模型GM（1，1）的预测过程及结果，并在此基础上进行马尔柯夫处理，建立灰色系统-马尔柯夫预测模型。其结果表明，后者预测精度明显提高，67％的样本误差可降至10％以内，92％的样本预测值误差低于30％。

Grey-Markov Predicting Model to Land Subsidence Evolvement

The rate of land subsidence has always been dynamic waveform,taking great difficult to the prediction.In this paper,the calculation and prediction of GM(1,1) to bench marks' land subsidence in Tianjin City may be improved by Markov Process.the final results show that the accuracy of Grey- Markov Predicting Model has been greatly more than GM(1,1),which 67% of the samples' predicting relative errors could be reduced to less than 10%,92% of the samples' predicting relative errors has been less than 30%.