基于机器学习算法的基坑沉降变形组合预测模型OA

In order to overcome the limitations of a single machine learning model and achieve high-precision pre-diction of foundation pit settlement & deformation,a combined prediction model for foundation pit settlement defor-mation is established by entropy weight method based on BP neural network model,convolutional neural network(CNN)model,support vector regression(SVR)model and long short-term memory(LSTM)model.Based on the moni-toring data of engineering examples,the prediction results of different models are evaluated.The results show that both single machine learning models and the combined model have high prediction accuracy;Based on the principle of information entropy,the weights of BP neural network model,CNN model,SVR model and LSTM model in the combined model are 0.253 2,0.265 2,0.236 1 and 0.245 5,respectively;The prediction accuracy of the combined model for the test set is the highest,the absolute coefficient R2 is 0.92,and the average relative error is 2.1％.The research results can provide reference for improving the prediction accuracy of foundation pit settlement deformation and ensuring the safety of foundation pit engineering.