Support Vector Machine Model Based on Reinforced Leader Decision Grey Wolf Optimization Algorithm

To address the problem of weak parameter optimization capability in support vector machine (SVM) models that leads to inaccurate prediction results, a novel approach is proposed based on the Reinforced Leader Decision Grey Wolf Optimization algorithm (DGWO) for SVM.This algorithm employs the MSE, MAE, and MAPE as performance indicators to evaluate the prediction performance of DGWO.Leveraging the global optimization capability of the DGWO algorithm, the positions of the wolf pack within the target range are used to represent the values of the SVM penalty factor (PF) and kernel function parameters (KFP).By performing a limited number of iterations and updating the positions using the reinforcement leader decision mechanism of the DGWO algorithm, the optimal PF and KFP for the SVM model are obtained.Experimental results show that the algorithm can find suitable parameters.In comparison with existing algorithms, the proposed approach exhibits better prediction performance and significantly reduces the optimization time.