A Novel Deformation Prediction Model for Mine Slope Surface Using Meteorological Factors Based on Kernel Extreme Learning Machine

Abstract

Extreme learning machine (ELM), as an emergent technique for training feed-forward neural networks, has shown good performance on various learning domains. This work evaluates the effectiveness of a new Gaussian kernel function-based extreme learning machine (KELM) algorithm for the deformation prediction of mine slope surface utilizing various kinds of meteorological influence factor data including the temperature, atmospheric pressure, cumulative rainfall, relative humidity and refractive index of the mining slope. The KELM model was applied to the deformation of Anjialing diggings, which is an open-pit mine of the China Coal PingShuo Group Co., Ltd. in China. The prediction performance on real data suggests that the proposed KELM model can effectively express the non-linear relationship between the mine open-pit slope surface deformation and various kinds of meteorological influence factors. The prediction results are compared with the most widely used algorithms – Support vector machine (SVM) and back-propagation neural networks (BP NN) in terms of the ease of use ( for example, the number of user-defined parameters), regression accuracy and computation cost. The comparison shows that the new algorithm is similar to SVM and BP NN but more accurate, and has notable lower computational cost and stronger generalization ability.