Evaluation method of slope instability space-time probability based on machine learning

The invention discloses a side slope instability space-time probability evaluation method based on machine learning, which has better prediction precision and generalization ability, and comprises the following steps: Step 100, based on a Bootstrap algorithm, a GRU algorithm and a Kriging algorithm, establishing a BGK side slope displacement space-time uncertainty prediction model, and outputting a side slope displacement space-time uncertainty prediction result; step 200, excavating a slope displacement space-time uncertainty prediction result based on a reliability theory, establishing a slope instability space-time probability evaluation model, and outputting a slope instability space-time probability interval; and Step 300, constructing a slope full-section displacement-instability probability binary coupling analysis index DP based on the most disadvantageous principle, and judging the overall safety of the slope according to the value of the DP. 本发明公开了一种具有较好的预测精度和泛化能力的基于机器学习的边坡失稳时空概率的评估方法，包括以下步骤：Step100