基于BP算法的林分材种出材率模型研究

A Study on the Stand Merchant Radio Prediction Model Based on BP Algorithm

通过对伐区设计资料，及实际生产码单数据进行学习，确定以平均胸径、平均树高、保留密度、蓄积量为输入神经元，分析了影响BP网络学习效率和预测精度的影响因素，主要从隐含层神经元数量、训练数、隐舍层激励函数、学习样本数量几个方面对材种出材率预测BP网络模型进行了优化，确定了林分经验材种出材率预测人工神经网络模型。

Design information through the cutting area, and the actual production code to study a single data to determine the average diameter at breast height, average tree height, retention density, accumulation of input neurons, and analyzed the impact of BP network learning efficiency and prediction accuracy of the impact of factors, primarily From the number of hidden layer neurons, training the number of hidden layer activation function, number of samples to study several aspects of the timber timber rate is forecast to grow BP network model has been optimized to determine the experience of the forest timber timber rate is forecast to grow artificial neural network model. Experience as a stand Merchantable Volume Table for the rate to provide a new thinking and methods.