Abstract:To accurately and quickly analyze the stability of open-pit mine slopes, this paper proposes Adaboost-Convolutional Neural Network (AdaBoost-CNN) to establish the nonlinear relationship between open-pit mine slope stability and influencing factors. AdaBoost-CNN is a new machine learning method that combines AdaBoost and CNN. The core idea is to combine the feature extraction capabilities of CNN with the ability of AdaBoost to process unbalanced data, which has the advantages of high reliability, high accuracy, and less training time. Adaboost-CNN uses migration learning, which not only eliminates the limitation of CNN that require a large number of training samples but also solves the problem of reducing actual performance in the serialization process of AdaBoost. This paper uses BP neural network, Support Vctor Mchine, CNN, and AdaBoost-CNN to model and analyze the measured engineering data. By comparing RMSE and RPE, it is found that AdaBoost-CNN has the highest prediction accuracy and the strongest model generalization ability. The results show that AdaBoost-CNN can predict slope stability more accurately, and is a reliable tool for slope stability prediction.