马尔科夫方法修正的灰色模型在吉林省粮食产量预测中的应用

粮食生产是国民经济重要的组成部分,粮食生产的波动必然会引发整个国民经济的波动。因此人们在努力提高粮食产量的同时,也期望知道未来一段时间粮食产量的变化情况,以便为科学决策提供依据。基于吉林省1949~2008年粮食总产量数据,采用灰色GM(1,1)预测模型动态模拟该省粮食产量变化态势,并运用马尔柯夫状态转移矩阵对灰色GM(1,1)模型的模拟结果进行修正,以提高粮食产量预测精度。结果表明,马尔柯夫方法修正的灰色模型能够大大提高粮食产量的模拟精度,模型修正后的模拟产量的相对误差较之修正前下降了0.10(由0.19下降到0.09),将灰色GM(1,1)模型和马尔柯夫状态转移矩阵相结合用于粮食产量预测可以取得较好的效果。预测结果表明未来10a吉林省将增产粮食100亿kg,增产潜力巨大。

Application of Grey Model Modified by Markov Method in the Yield Prediction of Jilin Province

Grain production is the important part of national economy development, whose fluctuation will make much effect on national economy safety. Therefore, people also need to know the grain production changes ahead of crop harvest period while they make an effort to increase grain production, for the purpose of providing the basis of scientific decision-making. Based on the grain yield data from 1949 to 2008 of Jilin province, Grey Prediction Model (1, 1) was utilized for dynamic simulation of the yield changes, and Markova Transition Matrix was introduced to amend the simulated result to improve the yield predicted accuracy. The studied results indicated that, the accuracy of the amended simulating yield increased greatly compared with that simulated by original GM (1, 1) (the relative error decreased obviously from 0.19 to 0.09). It is feasible of using a combination of grey model GM (1, 1) and Markov Chain Method for simulated grain yield. It was found that the total grain yield of Jilin Province will increase 10 billion kilogram in the next 10 years, which showed the great potential of grain producing in Jilin province.