摘要
为了解决各种单一传统的预测方法中存在的问题,提出了一种BP神经网络最优组合的预测方法。将单一预测方法所得到的预测值作为BP神经网络的输入样本,相应历史数据的实际值作为样本的输出,经过样本训练达到期望精度,应用BP神经网络进行预测。通过对浙江省农机总动力需求预测,表明该方法比各种单一的预测方法都有更高的精度。
Aiming at solving the problems in each traditional forecasting method, a optimal mix forecas- ting method based on BP neural network was put forwards. Using the for sample inputs of the BP neural and the history factual values as the outputs,Expected precision could be reached by sample training,the forecast could be done by applying BP neural network. By having applied the new method to the prediction of the total power requirement of agricultural machinery in Zhejiang province, it showed that the method had achieved better forecasting results compared with other forecasting models.
出处
《农机化研究》
北大核心
2004年第3期162-164,共3页
Journal of Agricultural Mechanization Research
关键词
BP神经网络
预测方法
人工智能
农业机械
动力需求
artificial intelligence
BP nearal network
application
the total power of agricultural machin- ery
