摘要
鉴于单一预测模型和线性组合预测模型的局限性,在确定黑龙江省农机总动力单一预测模型的基础上,建立了基于BP神经网络的非线性农机总动力组合预测模型。误差分析表明,该非线性组合预测模型的拟合平均绝对百分误差为3.03%,低于一元线性回归模型、指数函数模型、灰色GM(1,1)模型和三次指数平滑模型的6.26%、4.65%、4.88%和3.72%;稍高于以误差平方和最小为原则构建的线性组合预测模型的2.86%。用2006~2008年黑龙江省农机总动力进行检验预测,结果表明该模型可以有效地提高农机总动力的预测精度,用该模型预测了黑龙江省2009~2015年农机总动力。预测结果表明,在未来几年黑龙江省农机总动力将保持快速增长趋势,到2015年将达到40 537 MW。
In view of the limitations in single prediction models and linear combined prediction model,nonlinear combined prediction model for total power of agricultural machinery was put forward on the basis of establishing single prediction models for total power of agricultural machinery in Heilongjiang province.The results of error analysis showed that mean absolute percent error of proposed nonlinear combined prediction model was 3.03%,which was lower than 6.26%,4.65%,4.88% and 3.72% of one-variable liner regression model,exponential model,GM(1,1)model and cubic exponent smooth model,and a little higher than 2.86% of the linear combined prediction model based on the minimum sum of error square.Predicting total power of agricultural machinery from 2006 to 2008 proved that this prediction model could efficiently improve prediction accuracy for total power of agricultural machinery.The total powers of agricultural machinery were predicted from 2009 to 2015 in Heilongjiang province.The prediction results showed that total power of agricultural machinery would maintain swift growth tendency in the future several years,it would be 40537MW in 2015.
出处
《农业机械学报》
EI
CAS
CSCD
北大核心
2010年第6期87-92,共6页
Transactions of the Chinese Society for Agricultural Machinery
关键词
农机总动力
非线性组合预测
BP神经网络
Total power of agricultural machinery
Nonlinear combined prediction
BP neural network