<![CDATA[This study aims to systematically optimize the hyperparameters of an Artificial Neural Network (ANN) using an algorithm-based approach to obtain the best configuration, as well as to compare the air temperature prediction accuracy of ANN with other machine learning algorithms. Feature selection was performed using a physical approach to identify variables that significantly influence air temperature, while ANN hyperparameter optimization was conducted using the Random Search method. The ANN model was compared with Linear Regression (LR), Random Forest (RF), and Support Vector Regression (SVR) in predicting air temperature. Model performance was evaluated using Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and the coefficient of determination (R²). The results indicate that the ANN model generally achieved lower MSE, RMSE, MAE, and MAPE values and higher R² values compared to the other methods across four study locations: Banyuwangi, Juanda, Sangkapura, and Pasuruan. These findings demonstrate that ANN with properly optimized hyperparameters can provide more accurate air temperature predictions.. Keywords: Artificial Neural Network, East Java, Machine Learning, Air Temperature]]>
Copyrights © 2026
