基于无偏灰色马尔科夫链的稻瘟病预测研究

Study on Predictron of Rice Blast Area Based on the Unbiased GM(1,1) Model

采用五点滑动法优化原始数据,建立新的灰色马尔科夫预测模型。在分析利用湖北省恩施州1995～2004年稻瘟病历史资料的基础上,分别应用五点滑动优化无偏和无偏GM(1,1)模型预测2005～2010年稻瘟病发生面积。预测结果表明:五点滑动优化无偏GM(1,1)模型预测精度较高,应用于稻瘟病预测切实可行。结合五点滑动优化无偏GM(1,1)模型预测结果,笔者给恩施州2010年稻瘟病的防治提供了政策建议。

To create a new prediction model,the unbiased GM（1,1） model was optimized by the five-point sliding method in this paper.Then,based on analyzing and wtilizing the occurrence areas of rice blast in Enshi city from 1995 to 2004,the new model and unbiased GM（1,1） model was applied to predict the occurrence areas from 2005 to 2010 and compared.Predicting Outcomes showed that the forecasting precision of five-point unbiased sliding optimization GM（1,1） model was higher than the unbiased GM（1,1） model.Finally,combined with the forecast results,some suggestion for Enshi city in the prevention and control of rice blast in 2010 were provided.