Answers are grounded in this paper and cite their sources.
Accurately predicting dam slope displacement is a crucial prerequisite for ensuring the operational safety of hydraulic structures and preventing potential landslide disasters. However, existing prediction methods remain inadequate in the collaborative realization of temporal feature representation and high-precision regression, and their performance is especially limited under conditions of high dimensionality, noise interference, and small sample sizes. To address this issue, this paper proposes a hybrid prediction model that integrates a Convolutional Neural Network (CNN) and an AdaBoost-enhanced Relevance Vector Machine (RVM). First, the displacement time series is reconstructed via a sliding window to construct feature samples suitable for CNN input, thereby extracting local correlation structures and multi-scale features from the time series. Subsequently, the extracted features are fed into the AdaBoost–RVM model for regression prediction, so as to improve the prediction accuracy and generalization capability. The model is validated using monitoring data from a dam slope. The results demonstrate that, compared with various benchmark models, the proposed method achieves superior performance in terms of MAE, RMSE, MAPE, and R², and can well capture the displacement variation trend with stable predictions. The proposed method can provide an effective approach for the deformation analysis of dam slopes.
No takes yet. Share an insight, caveat, or question.
Li et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: