Comparative Analysis of Data-Driven Techniques to Predict Heating and Cooling Energy Requirements of Poultry Buildings

Many models have been developed to predict the energy consumption of various building types, including residential, office, institutional, educational, and commercial buildings. However, date, no designed specifically poultry buildings’ consumption. To address this information gap, study integrated data-driven techniques, artificial neural networks (ANN), support vector regressions (SVR), random forest (RF), into a physical model buildings in different climatic zones Turkey. The following statistical indices were employed evaluate model’s effectiveness: Root mean square error (RMSE), absolute percentage (MAPE), coefficient determination (R2). calculated predicted values heating cooling loads also compared using visualization techniques. results indicated that RF was most accurate during testing period according RMSE (0.695 6.514 kWh), MAPE (3.328 2.624%), R2 (0.990 0.996) for loads, respectively. Overall, offers simple decision-support tool estimate requirements weather conditions.