A crop recommendation using improved expert-guided constraint aware differential evolution and CatBoost surrogate...

Area of Science:

• Agricultural Science
• Computer Science
• Environmental Science

Background:

• Sustainable agriculture planning requires intelligent decision-support systems.
• Managing complex agro-environmental factors and multi-criteria interactions is crucial.
• Existing systems often lack scenario-awareness and adaptability.

Purpose of the Study:

• To develop a scenario-aware crop recommendation system.
• To integrate advanced optimization and machine learning for crop suitability assessment.
• To facilitate adaptive decision-making for sustainable agriculture.

Main Methods:

• Coupling Improved Expert-guided Constraint-Aware Differential Evolution (ECA-DE) with CatBoost (Categorical Boosting) surrogate learning.
• Assessing crop suitability based on agro-environmental factors (soil type, pH, temperature, humidity, water needs, duration, water source).
• Evaluating performance under balanced, drought, and short-season scenarios.

Main Results:

• Achieved high prediction accuracy (R² = 0.9975, MAE = 0.0043, RMSE = 0.0056), indicating near-perfect agreement with expert judgments.
• Improved ECA-DE significantly reduced computation time and memory usage while maintaining optimization quality.
• Demonstrated accurate representation of trade-offs in suitability, water requirement, and crop duration for adaptive recommendations.

Conclusions:

• The proposed framework offers a precise, interpretable, and resource-efficient tool for smart agricultural decision-making.
• The system effectively supports adaptive crop recommendations under diverse environmental conditions.
• This approach enhances sustainable crop planning and resource management.