Application Research of Support Vector Regression in Coal Mine Ground-Water-Level Forecasting

The forecast of the mine Ground-water-level is an issue with many influencing factors, highly non-linear and temporal series. SVR (Support Vector Regression) is applied to forecast Coal Mine Ground-water-level in this paper. Appropriate kernel function and parameters are chosen based on the analysis to SVR regression algorithm. This paper proposes the Forecasting Model of Coal Mine Ground-water-level basing on SVR regression algorithm and determines the forecast of the input factor and the output factor according to the physical geography and the hydrology geology situation of the chosen mining area. The numerical test results show that the forecast results have compatibility with the actual measurement result. We verify that the forecast model of Coal Mine Ground-water-level has effect, and provide a new effective method to the Forecasting of Coal Mine Ground-water-level.