Agri Assist: An AI Integrated Farmer Assistant

In this paper, an AI Agriculture Assistant integrating a crop recommendation system and an agricultural query chatbot is developed to provide guidance on crop cultivation. Based on this, the crop recommendation system uses a stack ensemble model comprising Random Forest and Gradient Boosting, and an accuracy of 99.32% and F1-Score of 99.26%. Here, word embeddings are performed through FastText which allows for a quick response time (0.0244 seconds) and a cosine similarity of 0.88629 for the chatbot. With RSA encryption for securing user data, we ensure secure communication. It uses the critical elements of the soil composition, weather and crop performance in formulating tailored agricultural guidance. Back end communication is done through Flask, and the entire system was deployed in this manner with a web interface to drive usage and for real time data responses. This entire application helps the farmers in decision making regarding crops cultivation and management through accurate and secure assistance.