Predictive Analytics and Decision Intelligence for Climate-Resilient Agritech Systems

Abstract

In this study, an architecture of climate-resilient crop recommendation is proposed that uses various machine learning models along with Decision Intelligence to ensure reliable crop recommendation. Five models, namely Logistic Regression, Decision Tree, Random Forest, SVM, and KNN, were used to predict the best crop based on soil and environmental factors. With an accuracy of 99.32%, F1-score of 0.993, and AUC of 0.9999, Random Forest had the best performance. Meanwhile, the Decision Intelligence layer used majority voting to aggregate predictions in order to reconcile contradictory outputs and generate recommendations based on consensus. Excellent crop class discrimination and few misclassifications were validated by confusion matrix and ROC studies. In order to enable well-informed decision-making for climate-resilient agriculture, the suggested framework shows how combining predictive analytics with decision intelligence can produce extremely accurate, comprehensible, and useful crop recommendations.