销量预测的相对误差马尔柯夫链模糊修正模型研究

在销量预测的改进型灰色预测GM(1,1)模型中,针对传统的相对误差修正的马尔柯夫链预测法,在相对误差微小变化的情况下,存在状态跃变的不合理性,提出应用马尔柯夫链模糊模型的方法.并开展实证研究,得出改进的销量预测值为103.5095万箱的结论.

The Research on Relative Error in Sales Volume Prediction and the Markov Chain Fuzzy Assistant Model

The paper proposes to apply Markov chain fuzzy model,in the modified gray prediction GM(1,1) model of sales prediction,based on the traditional relative error correction method in Markov Prediction,in which on the condition of the errors' tiny changes,there is irrational state of the jump.The paper also establishes state of fuzzy sets in domain,calculates the transfer ratio in fuzzy state,and forms the fuzzy transition possibility matrix,which aims to raise the precision in the prediction by modifying the relative error.By empirical study on the subject,the paper gets the conclusion:the modified sales volume prediction number is 103.5095