Accurate prediction of earthquake—induced building damage is essential for timely disaster response and effective risk mitigation. This study explores a machine learning (ML)-based classification approach using data from the 2015 Gorkha, Nepal earthquake, with a specific focus on reinforced concrete (RC) structures. The original dataset from the 2015 Nepal earthquake contained 762,094 building entries across 127 variables describing structural, functional, and contextual characteristics. Three ensemble ML modelsGradient Boosting Machine (GBM), Extreme Gradient Boosting (XGBoost), and Light Gradient Boosting Machine (LightGBM) were trained and tested on both the full dataset and a filtered RC-only subset. Two target variables were considered: a three-class variable (damage_class) and the original five-level damage grade (damage_grade). To address class imbalance, oversampling and undersampling techniques were applied, and model performance was evaluated using accuracy and F1 scores. The results showed that LightGBM consistently outperformed the other models, especially when oversampling was applied. For the RC dataset, LightGBM achieved up to 98% accuracy for damage_class and 93% accuracy for damage_grade, along with high F1 scores ranging between 0.84 and 1.00 across all classes. Feature importance analysis revealed that structural characteristics such as building area, age, and height were the most influential predictors of damage. These findings highlight the value of building-type-specific modeling combined with class balancing techniques to improve the reliability and generalizability of ML-based earthquake damage prediction.

This study aims to model the temperature distribution in a single fin subjected to steady one-dimensional heat conduction with nonlinear thermal behavior. For the modeling and solution of the problem, the Physics-Informed Neural Networks (PINNs) architecture was used. The temperature-dependent heat conduction problem and the nonlinear boundary conditions of this problem were formulated with a differential equation. With the help of the PINN architecture, the loss function was minimized in order to reduce the difference between the true value and the predicted value. During this minimization process, the PINN architecture was forced to be consistent with the physical laws. The results obtained after training the PINN architecture exhibit successful performance in terms of accuracy and reliability when compared with the results in the literature. These findings highlight the potential of PINNs as a powerful alternative to conventional methods for solving complex nonlinear heat conduction problems.

In this study, boosting machine learning techniques were employed to estimate the concentration of manganese in the potable water of Yuvacık Dam and to predict the days that surpassed and fell below the threshold value of 0.05 mg/L in Turkey. We conducted both regression and classification analyses for the same issue. We also implemented sampling methods when the data distribution in the classification task became imbalanced. We obtained daily measurements for approximately seven years, from 2004 to 2011, to build the dataset, which consisted of seven columns in total. While the XGBoost algorithm forecasted the manganese content in potable water with a mean absolute error of 0.0055, the Light GBM algorithm predicted the days with elevated manganese levels with an accuracy of 0.97. The models’ high predictions allow us to adjust the frequency of frequent water sampling and lab analysis and take prompt action during water filtration processes.