关键词: 开采沉陷预测, BP神经网络, 粒子群算法, 自适应增强算法

Abstract: Aiming at the complex nonlinear relationship between mining subsidence and multiple influencing factors, an Adaboost strong prediction model (Adaboost-PSO-BP model) based on particle swarm optimization to optimize BP neural network is proposed. The prediction accuracy is improved, and the mean of the average relative error is optimized. The results shows that the strong prediction model combines the characteristics of the Adaboost algorithm with a large prediction error and the particle swarm algorithm to optimize the weights and thresholds of the neural network, which achieves the purpose of "optimizing the best" of the strong predictor, which confirms the Adaboost-PSO-BP The feasibility and practicability of strong prediction model in the prediction of mining subsidence.