Enhancing Smart Greenhouse Management System Using Hybrid Machine Learning Model

This research introduces a new hybrid machine-learning model for optimizing smart greenhouse operations in response to worldwide food production demand. The model combines Random Forest (RF) and Support Vector Machine (SVM) algorithms to efficiently monitor and control important environmental parameters like temperature, humidity, water, and soil nutrient levels. With a large dataset of 37,923 data points and 13 essential features, the hybrid model uses RF for regression tasks like forecasting continuous environmental variables and SVM for binary classification tasks like predicting actuator states. Several conventional machine-learning techniques, such as Logistic Regression, Random Forest, Decision Tree, Naïve Bayes, SVM, and K-Nearest Neighbors (KNN), were used to compare the hybrid model’s performance. The hybrid model outperformed the others with an accuracy of 99.70%, precision of 99.75%, a recall of 99.71%, and an F1 score of 99.73%. These indicators highlight how well the model can improve greenhouse management by making accurate and dependable forecasts. This study demonstrates how smart agriculture may be upgraded using cutting-edge machine-learning techniques, providing a reliable solution for sustainable and effective greenhouse operations.