Akademik Veri Yönetim Sistemi
EFES INTERNATIONAL SCIENTIFIC RESEARCH AND INNOVATION CONGRESS-III, İzmir, Türkiye, 13 Eylül - 15 Ekim 2025, ss.818-819, (Özet Bildiri)
This study investigated the physiological responses of ‘Macit 55’ and ‘Akyüz’ hybrid chestnut (Castanea spp.) cultivars to waterlogging stress. The experiment consisted of a control group (under non-waterlogged conditions), a waterlogged group without salicylic acid (SA), and a waterlogged group treated with salicylic acid (SA). This study was conducted for five weeks in 2025 at the greenhouse unit of Ondokuz Mayıs University, Faculty of Agriculture. Plant height and chlorophyll content (SPAD) were measured on June 30, July 14, and July 28. Relative water content (RWC) and electrolyte leakage (EL) were also determined based on the collected samples. Waterlogging caused significant reductions in plant growth, RWC, and SPAD, while EL increased, indicating cellular membrane damage. Salicylic acid (SA) decreased some adverse effects of waterlogging by maintaining higher relative water content (RWC) and SPAD values. In addition to physiological evaluations, machine learning models were applied to predict RWC and EL. Extreme Gradient Boosting (XGBoost), Multilayer Perceptron (MLP), Gaussian Process Regression (GPR), and Random Forest (RF) MLP and ML models were tested. Gaussian Process Regression (GPR) and Random Forest (RF) provided the most accurate RWC predictions, achieving an R² of 0.93 and an RMSE between 2.67 and 2.68. In the case of EL, GPR outperformed the other models with the highest accuracy (R² = 0.96, RMSE = 2.60). Machine learning approaches effectively captured the nonlinear relationships among physiological traits under stress, demonstrating their potential as predictive tools in plant stress research. These results indicate that salicylic acid (SA) is important in improving chestnut waterlogging tolerance. Moreover, combining physiological data and machine learning models provides a practical approach for evaluating plant stress responses.