基于小波神经网络的加工番茄产量预测模型

新疆大学电气工程学院,乌鲁木齐 830047

计算机神经网络; 小波分析; BP神经网络; 加工蕃茄; 小波神经网络; 产量预测模型

A wavelet neural network model for processing tomato yield forecasting
Chen Yujia and Jiang Bo

School of Electrical Engineering, Xinjiang University, Urumqi 830047, P.R.China

computer neural network; wavelet analysis; back propagation(BP)neural network; wavelet BP neural network; proressing tomato; yield forecasting model

DOI: 10.3724/SP.J.1249.2015.05546

备注

基于小波分析和BP(back propagation)神经网络,建立加工番茄产量预测的小波神经网络模型,为制定合理的种植规划和管理决策提供科学依据.通过对新疆某番茄基地的历史数据进行分析,以温度、灌水量、氮肥、磷肥和钾肥的投入量作为模型的输入,番茄产量作为输出,建立5-10-1的预测模型.实验结果表明,预测值与实际值之间的最大相对误差仅为0.23%,收敛速度和预测精度均优于BP神经网络,实现了加工番茄产量的有效预测.

We propose a tomato yield forecasting model based on wavelet and back propagation(BP)neural network. By analyzing historical data of a tomato production base in Xinjiang, we build a neural network model with 5-10-1 structure where the climate temperature, irrigation amount and applied amount of chemical fertilizers function as inputs, and the yield prediction of tomato production act as the output. Experimental results show that the maximum relative error between the predicted and real output is only 0.23%. The convergence speed and prediction accuracy of the wavelet BP network are better than those of the conventional neural networks, indicating that it is a more effective model to predict tomato yield.

·