Random Forest Models Outperform Radial Basis Networks for Dam Prediction

A study published in the journal AI in Civil Engineering evaluates the predictive capabilities of machine learning tools for monitoring the structural behavior of concrete arch dams. The research, led by A. M. Babadi, H. Mirzabozorg, and K. Baharan, utilizes the Karun-I dam as a specific case study to assess how different algorithms can forecast structural responses. The authors focus on established open-source machine learning frameworks, specifically CatBoost, XGBoost, LightGBM, and TensorFlow, which are categorized within the study as being based on Forest and Radial Basis Function Networks. The primary objective of this work is to determine which of these computational approaches offers the most accurate and efficient means of analyzing dam displacement and other structural indicators.

The results of the analysis demonstrate that Random Forest-based methods achieve superior prediction accuracy compared to Radial Basis Function Networks. In addition to accuracy, the study finds that these Forest-based models also offer greater computational efficiency.

A key component of the research involves feature importance evaluation, which identifies the specific environmental and operational factors that most significantly influence dam displacement. The analysis reveals that lake levels are the primary factor impacting the displacement of the Karun-I dam.

The research underscores the promising potential of machine learning in enhancing structural health monitoring for large dams. The study offers insights that contribute to the improvement of safety measures and operational efficiency in dam management.

The study is published as an open access article in Volume 4, article number 22 of the journal AI in Civil Engineering.

Subject of Research: Comparative evaluation of Random Forest and Radial Basis Function Networks for predicting concrete arch dam structural behavior

Article Title: Predictive modeling of concrete arch dam behavior: evaluating the efficacy of Random Forest and Radial Basis Function Networks

Article References: Babadi, A. M., Mirzabozorg, H., & Baharan, K. (2025). Predictive modeling of concrete arch dam behavior: evaluating the efficacy of Random Forest and Radial Basis Function Networks. AI in Civil Engineering, 4(1), Article 22. https://doi.org/10.1007/s43503-025-00071-9

Image Credits: AI Generated

DOI: 10.1007/s43503-025-00071-9

Keywords: concrete arch dam, machine learning, Random Forest, Radial Basis Function, structural health monitoring, Karun-I dam, civil engineering, predictive modeling, CatBoost, XGBoost, Predictive, modeling