Explainable Al-enabled Mixed-crop Recommendation Model using Hybrid Machine Learning and Ant Colony Optimization

Mixed-cropping are widely recognized for enhancing land-use efficiency, improving soil fertility and contributing to long-term agricultural sustainability. However, choosing optimal crop combinations tailored to specific farm conditions remains a challenging, particularly in developing regions.

To address this limitation, this study proposes a novel explainable artificial intelligence (XAI)-Driven hybrid machine learning and ant colony optimization (ACO) framework (XAI-HACO) for mixed-crop recommendation, integrating soil nutrient, weather variables. hybrid machine learning techniques such as random forest-extra trees (RF-ET), decision tree-C4.5 (DT-C4.5), extreme gradient boosting-gradient boosting (XGBoost-GBoost), quadratic discriminant analysis-linear discriminant analysis (QDA-LDA) and support vector machine-stochastic gradient decent (SVM-SGD) were developed and assessed using an indigenous mixed-crop dataset of Andhra Pradesh, India. The models performance was assessed using accuracy, precision, recall, F1-score, confusion matrix and ROC-Curve, while ACO was employed to optimize feature selection and model hyperparameters.

The investigational results show that the proposed RF-ET hybrid model achieved superior performance 95.91% accuracy, precision at 95.08%, recall at 95.91% and F1-score at 95.49%. These results show that the proposed XAI-HACO framework offers a reliable, transparent, interpretability and data-driven decision support tool for farmers and agricultural stakeholders, facilitating informed selection of suitable mixed-cropping systems.