改进灰色-马尔科夫模型在年降水量预测中的应用研究

Application of Improved Grey-markov Model in Annual Precipitation Prediction

用三点滑动法对灰色GM(1,1)模型优化,用模糊聚类FCM算法对马尔科夫模型状态分组阶段进行改进,用多种步长的马尔可夫链加权算法来预测年降水量。将上述改进后的灰色模型与马尔科夫模型结合,对地区年降水量进行预测。结果表明,改进后的灰色-马尔科夫模型既提高了马尔科夫模型对预测波动性降水数据的准确性,又充分利用了灰色系统预测结果精度高的优势,提高了预测精度。

In this paper,three sliding method is used to optimize grey GM(1,1)model,fuzzy clustering FCM algorithm is used to improve the divide of Markov model group stage,and the weighted Markov chain algorithm with a variety of step length is used to predict the annual precipitation.Then the improved grey model and Markov model is combined to forecast the regional annual precipitation.The results show that the improved grey-Markov model can not only improve the changeful precipitation data prediction accuracy of Markov model,but also make full use of the grey system's advantage of high precision,so improves the prediction accuracy.