Expert System for Disease in Cotton Plant Using Task-Oriented Dialogue
B Jothika, Laxmi B Rananavare · 2024
Crop production is significantly hampered by plant diseases, necessitating early and accurate identification to mitigate losses. Despite advancements in machine learning (ML), image processing, and deep learning (DL), agricultural disease management remains challenged by the complexity and opacity of current AI methodologies. This study introduces an expert system developed using SWI-Prolog to identify plant diseases based on user queries. The system analyzes environmental factors including temperature, humidity, rainfall, evaporation, sunshine, and pest values. Knowledge extraction from a comprehensive dataset and the application of predefined rules enable the system to match user-inputted values against these rules for disease identification. The system demonstrates effective disease identification and provides detailed explanations, enhancing transparency. The integration of a Knowledge Base System (KBS) mimics human expert reasoning, further improving the system’s problem-solving capabilities. This approach significantly improves the transparency and efficacy of agricultural disease management practices.