Enhancing Crop Recommendations with Ensemble Learning and Explainable AI
Abstract
The classification technologies of AI can provide precision crop recommendations for modern agriculture, and play a significant role in improving agricultural production. By combing the relevant literature, this study proposed a new ensemble learning model based on comprehending the highly accurate classification results of ensemble models used in agricultural crop recommendation with AI and IoT. Then, the historical data is used to train and test while considering as many factors as possible from previous studies. The proposed model has the best accuracy at 99.32%, which is higher than any single classification model used in its ensemble process and the models in previous studies. Meanwhile, in terms of precision, recall and F1 score, the proposed model is superior to the single classification model utilized in its integration, which shows the great reliability. Two XAI techniques (SHAP and LIME) were used to interpret the model classification results, quantifying the contribution of each feature to model prediction, improving system transparency and thereby enhancing intelligibility for farmers.