Ground subsidence prediction method and device based on machine learning

The invention belongs to the technical field of geological disaster prevention, and relates to a land subsidence prediction method and device based on machine learning. Acquiring land subsidence rate information based on the InSAR and the land subsidence mark data; determining and acquiring land subsidence influence factors based on the drilling data, the mining quantity and the hydrogeological conditions; and constructing a land subsidence prediction model by using a machine learning algorithm, and quantifying the contribution degree of the land subsidence influence factors to realize land subsidence prediction. According to the method, the land subsidence value can be obtained only by inputting the main control factor value in model operation, the method has the characteristics of simplicity, rapidness, convenience, accuracy and the like, time and economic cost are saved for urban disaster prevention and control, and a scientific basis is provided for urban planning. 本发明属于地质灾害预防技术领域，涉及一种基于机器学习的地面沉降预测方法及装置。基