Evaluating Machine Learning Models for Greenhouse Temperature Prediction

Greenhouse management plays a crucial role in modern precision agriculture, requiring accurate prediction of environmental conditions, particularly temperature, to optimize plant growth. Traditional climate control methods rely on manual interventions, which are often inefficient and inconsistent. This study evaluates the performance of various machine learning models, including Random Forest, XGBoost, Support Vector Machines, Gradient Boosting, Artificial Neural Networks, and Long Short-Term Memory networks, to predict greenhouse temperature and analyze the influence of environmental factors such as humidity, CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> levels, and illuminance. A preprocessed dataset was used to train and evaluate these models using Mean Absolute Error and R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> Score. The results indicate that Random Forest and XGBoost achieved the highest accuracy, making them the most reliable models for temperature prediction, while Long Short-Term Memory models exhibited overfitting, limiting their effectiveness with the available dataset. Feature importance analysis further revealed that humidity and CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> levels significantly impact temperature variations, aligning with findings in previous research. These insights contribute to improving climate control strategies in precision agriculture. Future research should focus on optimizing deep learning approaches and integrating real-time sensor data for enhanced predictive accuracy.