关键词: 边坡变形预测　粒子群优化　支持向量机　机器学习

Abstract: Realizing the monitoring and analysis of mine slope deformation is of great significance for ensuring the safety of mine production.The traditional Quantum Particle Swarm Optimization (QPSO) algorithm has the advantages of strong glob al optimization ability and few control parameters,but it is prone to premature convergence,which affects the prediction accura cy.Therefore,an adaptive learning factor is introduced to dynamically adjust the particle search strategy,enhancing the global search ability of the QPSO algorithm and preventing premature convergence.At the same time,by combining the advantages of the Support Vector Machine (SVM) model in classification and regression problems,an integrated prediction model (IQPSO SVM model) based on the improved QPSO (IQPSO) algorithm and the SVM model is proposed.This model first optimizes the hyperparameters of the SVM model using the IQPSO algorithm,enabling the SVM model to better handle complex geological data;then,the optimized SVM model is applied to the prediction of mine slope deformation.Taking a certain mine slope as an example for verification analysis,the results show that this model has certain advantages in predicting the deformation of mine slopes compared to the SVM,Random Forest (RT) and PSO-SVM models.

Key words: slope deformation prediction,particle swarm optimization,support vector machine,machine learning