IoT-Assisted Heterogeneous Ensemble Learning Environment for Smart Farming

This research deals with machine learning, deep learning, and AI techniques to handle crop recommendation within a heterogeneous IoT-sensed environment. We have used several standalone machine learning classifiers, ensemble learning, meta classifiers as well as deep learning to showcase crop recommendation. In this article, we have considered the three performance parameters, i.e. precision, Recall, and Accuracy. It has been observed that the combination of SVM and CNN significantly outperforms as comparison to other methods. It obtained the result of 99% for crop yield prediction. We also explore the meta classifiers to predict suitable crops. These classifiers exhibit diverse performance metrics, highlighting their potential for enhancing crop recommendation systems. CATBoost obtained an accuracy of 99.15%, precision of 99.18%, and recall of 89% for crop recommendation. Apart from these, we also discuss the various sensors for creating a reliable IoT-assisted model for smart farming employing standalone machine learning algorithms, and meta classifiers in a Heterogeneous Ensemble Learning Environment (HELE), optimizing smart-farming predictions through enhanced preprocessing, feature extraction, and data balancing techniques, and utilizing deep learning for plant disease diagnosis.