In this article, a centrifuge shaking table model test of anchored stabilizing piles for strengthening landslides was established, and the dynamic response characteristics of the pile–anchor–slope under earthquake action were analyzed. On this basis, combined with the fuzzy gray relational analysis and the rank-sum ratio method, the correlation between the amplification of the acceleration response of the heterogenous slope and the dynamic response of the support structure was explored. Based on the obtained results, relevant suggestions for engineering design were proposed. The results showed that the seismic amplification of the complex soil–rock slope reinforced by the pile–anchor structure was not uniform and the amplification coefficient had strong variability. Among them, the amplification coefficient of the slope, dynamic earth pressure, and dynamic bending moment of the pile near the connection of the pile–anchor cable continued to increase; the correlation between the seismic amplification and the seismic behavior of the pile–anchor structure is different at different positions of the slope. The measurement points with a higher comprehensive ranking of correlation are mainly concentrated in the pile–anchor connection, the middle of the slope, and the high-angle soil–rock interface. It is related to the geometric characteristics of the model and the high seismic amplification of the slope; for the pile–anchor connection part and the high-angle soil–rock structure surface of the slope, the shock absorption measures and grouting strength of the anchor cable’s anchoring section should be considered in the engineering design.

Peak ground acceleration (PGA) prediction is of great significance in the seismic design of engineering structures. Machine learning is a new method to predict PGA and does have some advantages. To establish explainable prediction models of PGA, 3104 sets of uphole and downhole seismic records collected by the KiK-net in Japan were used. The feature combinations that make the models perform best were selected through feature selection. The peak bedrock acceleration (PBA), the predominant frequency (FP), the depth of the soil when the shear wave velocity reaches 800 m/s (D800), and the bedrock shear wave velocity (Bedrock Vs) were used as inputs to predict the PGA. The XGBoost (eXtreme Gradient Boosting), random forest, and decision tree models were established, and the prediction results were compared with the numerical simulation results The influence between the input features and the model prediction results were analyzed with the SHAP (SHapley Additive exPlanations) value. The results show that the R2 of the training dataset and testing dataset reach up to 0.945 and 0.915, respectively. On different site classifications and different PGA intervals, the prediction results of the XGBoost model are better than the random forest model and the decision tree model. Even if a non-integrated algorithm (decision tree model) is used, its prediction effect is better than the numerical simulation methods. The SHAP values of the three machine learning models have the same distribution and densities, and the influence of each feature on the prediction results is consistent with the existing empirical data, which shows the rationality of the machine learning models and provides reliable support for the prediction results.