eng Enviro Research Publishers Current Agriculture Research Journal 2347-4688 2321-9971 2026-10-21 14 3 27638 article Juhi Patel 1 Tejaskumar Bhatt 2 Faculty of Computer Applications and Information Technology, GLS University, India. Faculty of Engineering and Technology, GLS University, India The use of biological nutrient inputs depends on soil properties, crop requirements and climatic conditions. This study presents a machine-learning framework intended to support the classification of nutrient-input recommendations using soil, crop and climate attributes. The submitted study identifies Soil Health Card information, crop-related data and climate variables as its data sources and compares Random Forest (RF), Support Vector Machine (SVM), Gradient Boosting (GB) and Artificial Neural Network (ANN) classifiers. The original manuscript reports accuracies of 91.8% for RF, 89.6% for GB, 87.4% for ANN and 85.2% for SVM; these figures must be verified against the original experiment outputs and interpreted in light of the validation protocol. The final version must report the verified dataset size, exact target classes, label-generation procedure and train/test and cross-validation design. Since the described analysis uses secondary data and does not report field experiments, it does not establish that the framework increases yield, reduces chemical-fertilizer use or improves climate resilience. The framework is therefore presented as a preliminary classification and decision-support approach that requires independent agronomic validation. https://www.agriculturejournal.org/volume14number3/a-data-driven-micro-fertilizer-recommendation-framework-for-cleaner-and-climate-resilient-agriculture/ Agricultural decision support; Biological nutrient inputs; Biofertilizer recommendation; Climate variables; Machine learning; Soil health