Agrosphere: AI-Powered Personalized Scheme Navigator for Farmers

Abstract

Access to government welfare schemes remains a major challenge for farmers due to fragmented information, complex eligibility criteria and lack of personalized guidance. Despite the availability of numerous schemes related to subsidies, insurance and financial support, many eligible farmers fail to benefit from them because of low digital literacy, language barriers and dependence on intermediaries. To address these issues, this paper presents Agrosphere, an AI-powered personalized scheme recommendation system designed to assist farmers in identifying suitable government programs based on their individual profiles. The proposed system follows a modular three-tier architecture consisting of a React-based frontend, a Spring Boot backend and a Python-based machine learning microservice. It employs a hybrid approach that integrates a rule-based eligibility engine with machine learning algorithms such as Logistic Regression and K-Means clustering. The rule-based module ensures strict compliance with official eligibility criteria, while the machine learning component enhances personalization by analyzing patterns in farmer data and ranking schemes accordingly. The system processes key farmer attributes including landholding size, income category, crop type and geographic location to generate accurate and interpretable recommendations. Experimental evaluation demonstrates that the system efficiently filters irrelevant schemes and produces ranked outputs within acceptable response time for real-time applications. The proposed solution improves accessibility, reduces information asymmetry and minimizes dependency on manual assistance. Agrosphere highlights the potential of combining rule-based systems with machine learning to build scalable, transparent and user-friendly decision support platforms for e-governance, ultimately contributing to improved welfare delivery in the agricultural sector.