An improved least squares SVM with adaptive PSO for the prediction of coal spontaneous combustion

The problem of coal spontaneous combustion prediction is very complex, and there are many factors that affect the prediction results. In order to solve the issues of high dimension and redundancy among features and limited samples in the prediction of coal spontaneous combustion, this paper proposes a prediction algorithm of coal spontaneous combustion based on least squares support vector machine and adaptive particle swarm optimization (APSO-LSSVM). The adaptive PSO algorithm is used to solve the problems such as high computational complexity and slow calculation speed of the LS-SVM model for large-scale samples, so that it can always obtain the optimal solution, and its training speed and accuracy are improved. This method adjusts the inertia weight based on the convergence degree of group and the adaptive value of an individual for accelerating the training speed of swarm. After that, the improved PSO is used to iteratively solve the matrix equations in LS-SVM. APSO-LSSVM avoids the matrix inversion, saves the internal memory and obtains the optimum solution. The experiment results show that this method simplifies the training sample, accelerates the training speed, and also offers superior classification accuracy, fast convergence speed and good generalization ability.