CONRENET: A MODULAR CONTEXT-AWARE REMEDY RECOMMENDATION NETWORK FOR PLANT DISEASE MANAGEMENT

Abstract

 Plant diseases remain a major challenge for global agriculture, causing significant yield losses and economic damage each year. Although recent advances in artificial intelligence have improved disease detection accuracy, most existing systems are limited to diagnostic tasks and rely solely on image data, overlooking critical contextual factors such as crop stage, disease severity, and environmental conditions. To address these limitations, this paper proposes ConReNet (Context-Aware Remedy Recommendation Network), a lightweight, modular, end-to-end framework designed to generate crop- and context-specific remedy recommendations. ConReNet consists of three core components: (i) the Contextual Feature Interaction Encoder (CFIE), which captures complex dependencies among crop, disease, and environmental variables; (ii) the Sparse Contextual Fusion Module (SCFM), which efficiently aggregates heterogeneous contextual information; and (iii) the Context-Aware Remedy Ranking Network (CARRN), which produces ranked remedy recommendations. Synthetic datasets simulating realistic crop–disease–remedy interactions were constructed to evaluate the framework. Experimental results demonstrate that ConReNet effectively learns contextual relationships and delivers accurate recommendations, establishing its potential as a scalable solution for intelligent disease management. Its modular design enables future integration with image-based models and IoT data streams, making it suitable for real-world deployment in precision agriculture.