Forecast of inrushed water volume grade from coal floor based on support vector machine with particle swarm optimization

To solve the problem of penalty factor and kernel parameter of support vector machine (SVM) which will affect the forecast accuracy, the method was put forward to find the better parameter value by using particle swarm optimization (PSO) which can automatically search the parameters for SVM. Four indexes, including water pressure, the thickness of aquifuge, karst development degree, the fault scale, were selected as the factors influencing water inrush from coal floor, the actual cases of water inrush from coal floor in Northern China coalfield were taken as training samples, the PSO-SVM model for forecast of water inrush volume grade from coal floor was established and applied to test other cases. The application of the model indicated that the method can solve the small sample, nonlinear problem, and the results obtained is better in accordance with the practice. It is practical and effective in forecasting water inrush volume grade from coal floor.