Crop Recommendation System for Konkan Region of Maharashtra using Interpretable Machine-Learning with XAI

Machine Learning has proved beneficial in many domains positively impacting various aspects of our daily lives. Agriculture in India is a major occupation and nations economy is deeply affected by agriculture. This study suggests a Crop Recommendation System for the Konkan region of Maharashtra utilizing Machine Learning combined with Explainable Artificial Intelligence (XAI) in light of growing agricultural uncertainties and the urgent need for sustainable farming techniques. The Konkan region, a crucial agricultural zone in Maharashtra, is the direction of the dataset, which includes important agronomic metrics such as soil nutrients (nitrogen, potassium and phosphorus), pH level, temperature, humidity, and rainfall. The prediction performance of a number of supervised machine learning models, such as Random Forest, KNN, Support Vector Machine (SVM), Decision Tree and XGBoost, in suggesting the best crops based on these characteristics was assessed. SHAP (SHapley Additive exPlanations) values were used to illustrate feature contributions and model decisions in or-der to improve interpretability and confidence among local farmers and policymakers. LIME is used to explain the reason for suggestion of a specific crop for a given land. In addition to promoting openness, the incorporation of XAI enables stakeholders to make knowledgeable agronomic choices. By enhancing crop output, soil health, and financial returns, this technology can make a substantial contribution to precision agriculture in Maharashtra.