ANNADATTASAATHI: AN AI-POWERED DECISION SUPPORT ECOSYSTEM FOR PRECISION CROP SELECTION AND SOCIO ECONOMIC YIELD OPTIMIZATION

Abstract

Agriculture remains the backbone of India’s economy, employing nearly half of the population but contributing only about 18% to GDP. Most Indian farmers are smallholders managing fragmented land under uncertain climatic conditions. To address these challenges, this paper presents AnnadattaSaathi, an AI-powered decision support ecosystem designed for crop selection and socio-economic yield optimization. The proposed system combines a Random Forest–based recommendation model (achieving 99.31% accuracy in experimental evaluation) with a scalable MERN–Flask architecture to provide practical field-level recommendations. To improve usability and trust, the system integrates Explainable AI techniques such as SHAP and LIME along with a multilingual conversational interface powered by GPT-4o. Unlike many existing research prototypes, AnnadattaSaathi focuses on real-world deployment by combining machine learning, IoT sensing, and user-friendly advisory tools. The system aims to support Indian farmers with transparent, localized, and actionable insights for precision agriculture.