基于灰色理论与BP神经网络预测成都市生活垃圾产生量

摘要

According to historical data of municipal solid wastes quantity in central areas of Chengdu. The fitted value of his- torieal data was obtained by grey GM（ 1,1 ） prediction model with time as variable. The relations between gray correlation degree and municipal solid wastes quantity and its influencing factors were analyzed. Four factors were selected to establish a muhivari- able gray GM（ 1,5） prediction model and BP neural network model. The prediction accuracy of municipal solid wastes quantity with the two models was compared. BP neural network model whose prediction accuracy is the highest was adopted to forecast the waste quantity in following years. This paper provides theory basis for disposal plan of municipal solid wastes in the central city of Chengdu.%根据成都市中心城区垃圾产生量的历年数据，先用以时间为单变量的灰色GM（1，1）预测模型得到历年数据的拟合值，再分析垃圾产生量与其影响因素之间的灰色关联度，选出关联度最大的4个因素建立多变量的灰色GM（1，5）预测模型与BP神经网络模型，并对垃圾产生量的预测精确度进行了对比，用预测精度最高的BP神经网络模型对未来年份的垃圾产生量进行了预测，为成都市垃圾处理处置规划提供了理论依据。