A Hybrid IoT and Machine Learning Approach for Crop Recommendation Using a Voting Ensemble Model

Agriculture is a key component of a nation's economy. In the traditional method, farmers rely on experience to determine the ideal climate and environment for a given crop. Technology is advancing, and while the system depends on it to some extent, improvements are still required. This study presents an Internet of Things (IoT) based crop recommendation system that uses Machine learning (ML) to use the value that IoT devices collect from the soil. Initially, we utilized baseline ML classifier models like Support vector machine (SVM), K-Nearest Neighbor(KNN), Random forest (RF), and Random forest (NB).Then we employed an ensemble model to detect the suitable crops. We employed several IoT sensors in our research. The DHT11 sensor collects real-time environmental data, such as temperature and humidity. During crop recording, vital soil parameters like potassium and phosphorus—which are essential for crop growth—are measured using an NPK sensor, a pH sensor, and a rain sensor. These sensors can be seamlessly integrated by the IoT device, offering an effective way to collect crucial agricultural data. The Voting Base Ensemble performs exceptionally well in terms of precision, recall, and F1 score measures. The voting Base Ensemble algorithm improves metrics such as F-score, recall, and precision by up to 98%. A thorough evaluation of soil conditions is made possible by the integration of various sensors, and the application of an ensemble model enables accurate crop recommendations based on environmental factors.