Risk assessment of debris flow in abandoned mining area based on AHP-RBF neural network model

Based on the engineering background of three debris flow gullies in an abandoned iron mine area in East Guangdong Province, a radial basis neural network model （AHP-RBF neural network model for short） based on analytic hierarchy process is proposed. Based on the geotechnical investigation results of the site, the risk assessment factors of debris flow gullies in abandoned mining areas are assigned and neural network normalization training is conducted. Using MATLAB platform to carry out debris flow risk research in abandoned mining area. The results show that the evaluation model of AHP-RBF neural network has high computational accuracy, and can better bind the experience knowledge of experts to the network nodes in the way of weights, successfully simulate the thinking mode of experts, realize human-computer interaction, and avoid the excessive influence of the subjectivity of AHP on the risk assessment, so that the results are accurate and objective.