An Innovative Machine Learning System for Optimal Multi-Class Classification of Date Fruits by Leveraging Backward Feature Engineering to Promote Sustainable Agriculture

Sustainable agriculture is crucial for enhancing food security and optimizing resource use, particularly in high-value crops like date fruits. Accurate multi-class classification of date fruit varieties is essential for refining agricultural practices. Traditional classification methods frequently exhibit low accuracy and susceptibility to human errors. To address these limitations, this research introduces an innovative ML framework for optimal multi-class classification of date fruit varieties. By integrating feature engineering, the framework significantly improving model accuracy and performance, demonstrating superior classification efficiency. Advanced ML models, including Optimized Trees, Ensemble, KNN, and a novel Optimized SVM, are employed to tackle the complexities of multi-class classification. The proposed Optimized SVM achieved the highest classification accuracy. These findings showcase the potential of machine learning to improve sustainable agriculture by enhancing date fruit classification accuracy. This innovative approach promotes sustainable practices, supporting smart farming and resource optimization while contributing to sustainable development through improved multi-class classification performance.