基于改进灰色马尔科夫模型的木材需求量预测

Prediction of the wood demand by improved grey Markov mode

以木材需求量2005―2013年数据为原始数据,建立了传统灰色、无偏灰色和滑动无偏灰色模型,并分别以3种模型预测2014―2015年木材需求量。通过检验模型的拟合效果和对比预测的相对误差,选取预测效果最优的滑动无偏灰色模型为待修正模型。以滑动无偏灰色预测数据的相对误差分布为状态划分依据,考虑实际情况对一般马尔科夫修正方法进行优化,以优化的修正方法对预测数据进行修正。最终结果表明:当进行改进马尔科夫修正时,滑动无偏灰色预测模型的平均相对误差从6.44%降至1.49%,预测误差减少4.95%,预测精度明显提升,能够为我国未来木材需求量的准确预测提供可靠的理论依据。

Grey model was established, unbiased grey model and sliding unbiased model by taking the Timber demand from 2005 to 2013 as the original data, timber demands in 2014-2015 were predicted respectively. Through comparing the relative error of prediction and the fitting effect, the optimal prediction model-sliding unbiased grey-forecasting model was selected as correction model. With taking relative error of the sliding unbiased grey prediction as classification basis, the general Markov method is optimized considering the actual situation and the optimized Markov was applied to fitting forecast data. Final results show that the improved Markov combined with the sliding unbiased grey forecasting model of average relative error was from 6.44% to 1.49%, and 4.95% less prediction error, the prediction accuracy is improved significantly, can provide reliable theoretical basis for accurate prediction of wood demand in future.