人工神经网络在杉木产区划分中的应用研究

本文运用人工神经网络理论，对福建省的杉木产区划分进行研究．用具有S－型特征函数的B－P网络对每一产区的标准样本进行训练，由训练结束后的权植及其待测报样本的因子测值，计算出网络输出值，作出杉木产区分类预报．研究结果表明：人工神经网络用于杉木产区的划分具有科学性、可行性和实用性。

Study on Application of Artificial Neural Network to Classification of Chinese Fir Growing Area

The theory of artificial neural network was applicated to study the classification of Chinese fir growing area in Fujian poovince. The training on standard sample of each Chinese fir growing area was performed by using B-P network with sigmoid function. The forecasting of classification was made for Chinese fir growing area according to the calculated outputs of network in the basis of the trained weight and threshold values as well as the monitoring values of sample's factors. The results showed that the proposed method possesses objectivity, feasibility and practicality. in Chinese fir growing area forecast.