Based on Quantitative Assessment of Geological Structure Complexity by Support Vector Regression

In the study, support vector regression (SVR) combined with particle swarm optimization (PSO) for its parameter optimization, was proposed to establish a model to predict the gas content under five influential factors, including buried depth, bedrock depth and three quantitative assessment indexes (kd the fold structure complexity, kz the fold structure complexity, kq the inclined angle structure complexity). The prediction result demonstrated that the estimated error of the SVR model, such as mean absolute error (MAE) and mean absolute percentage error (MAPE), is less than that of backpropagation neural network (BPNN) and multivariate linear regression (MLR) model by applying identical training and test samples. This study suggests SVR can be used as a powerful approach to foresee the gas content in coal with different geological structures.