泥石流危险度评价的神经网络法

泥石流危险度的评价计算，对于泥石流治理及减灾防灾对策的确定具有重要意义，影响泥石流危险度的因素复杂且具有胡机和模糊特性，而神经网络的性能特征使其能适用于解决非线性的泥石流危险度评价问题，本建立了相应的评价泥石流危险度的神经网络模型，并利用具体的实例对网络进行训练和测试，计算分析表明，网络模型对于评价泥石流危险度有较好的适用性。

ARTIFICIAL NEURAL NETWORK METHOD FOR EVALUATING THE DANGEROUS DEGREE OF DEBRIS FLOWS

There are some senses in evaluation calculation of debris flows accrding to the dangerous degree for harnessing a debris flow and defining the way to reduce the effects of the debris flow. The factor which control and affect the dangerous degree of a debris flows are random and fuzzy variables, and neural network is suitably used to solve the non-linear problems because of it's character, such as the estimation of dangerous degree. The relevant neural network model is set up in the paper for the estimation of the dangerous degree of a debris flows, and the factual cases are used to train and exam the network. According to the computing and analyses, the neural network is applicable to estimate the dangerous degree of a debris flows.