摘要
文中首先用1976年到1995年的桐城县冬小麦单产,建立灰色GM(1,1)模型,再用随机过程理论的马尔可夫模型获得GM(1,1)模型在已知年份里的偏差规律(即偏差的转移概率矩阵),并且依照此规律对GM(1,1)模型结果进行修正,将由GM(1,1)模型预测的一个具体数值,修正成为区间和概率组成的预测范围,增加预测的可靠性。最后用灰色-马尔可夫模型外推预测1996年到2000年共5年的小麦单产。实验说明灰色-马尔可夫模型大大提高了预测精度,将预测结果表示为预测范围,更为准确地反映出粮食产量的走势。
GM ( 1,1 ) model is established by using winter wheat yield of Tong Cheng county (from 1976 to 1995). Then the deviation regularity of GM( 1,1 ) model from 1976 to 1995 is obtained by Markov model of stochastic process (i. e. transition probability matrix). To improve forecast reliability, the result of GM( 1,1 ) model is amended based on the regularity. The result of GM( 1,1 ) which is just a numerical value is changed to a range consisted of intervals and probabilities. At last,winter wheat yield of Tong Cheng county (from 1996 to 2000) are forecasted by grey-Markov model. In this way, the forecast precision is improved, and the trend of yield is reflected accurately.
出处
《计算机技术与发展》
2007年第6期191-193,196,共4页
Computer Technology and Development
基金
安徽省自然科学基金(050420208)
安徽省教育厅重点自然科学研究项目(2006KJ015A)