A multilingual smart farming system for real time agricultural decisions using machine learning

Despite being the backbone of the global economy, farmers have always struggled with challenges like what crop to plant, how much fertilizer to use, and how to detect the disease beforehand. These problems affect farmers and lead to various concerns, such as food safety, which will greatly impact sustainable agriculture practices. The old farming methods depend on emotion, intuition, feeling, and experience, reducing yield production and economic losses. This paper proposes a machine-learning-based Smart Farming and Advisory System (SFAS) to overcome the challenges. The proposed model helps farmers by providing real-time crop suggestions based on soil, climate, and market conditions. It also uses soil nutrient composition to provide custom-made fertilizer recommendations and uses the Deep Learning (DL) method to detect plant diseases by processing images. Additionally, the proposed system works on multi-language support, which enables farmers of diverse linguistic backgrounds to access agricultural insights in their native languages. This research seeks to bridge the knowledge gap in agriculture using Machine Learning (ML), the Internet of Things (IoT), and multilingual accessibility to make farmers self-reliant through data-driven decision-making and sustainable agriculture practices. The experimental results show that the proposed technique outperformed baseline ML techniques on considered performance matrices, such as accuracy, sensitivity, specificity, etc.