Prediction of Domestic Waste Output Based on Nonlinear Autoregressive Neural Network Model
In this paper,a prediction model for the domestic waste output was established to better predict the domestic waste output,which can help building a plan for domestic waste disposal facilities and a flexible collection and transportation system.By using the non-linear auto-aggressive neutral network(NAR)and testing the model parameters such as delay order and the number of hidden layer neurons,a prediction model based on historical time series of domestic waste output was built.The results showed:The NAR neural network time series model had good predictability for the domestic waste output in Beijing.When the delay order was 5 and the number of hidden neurons was 10,the model achieved an r-value of 0.9717,a mean absolute percentage error of 3.385％,and a root mean square error of 5051.831 t/w.The residual sequence also passed the non-autocor-relation test,indicating a good prediction performance.The conclusion also indicated that NAR neural network model can be used for nonlinear autoregressive prediction of domestic waste output,without considering the availability of other related influencing factors,and had specific convenience and practical application significance.
国家科技期刊平台
NSTL
万方数据
