Expert system for diagnosis pests and diseases of the rice plant using forward chaining and certainty factor method

Abstract

Pests and diseases are one of the main factors that affect the low level of rice plant productivity. The symptoms in the infected rice plant are sometimes difficult to identify because they often shows the similar signs or characteristics so that only the experts who can identity them correctly. The infected rice plant actually can be identified since the beginning stage of planting until harvest time. So by knowing the symptoms in the early stage of the rice plant growth some preventif actions then can be done. Identifying pests and diseases of rice plant needs skills, experiences, and knowledge and should be done fast and accurate because the pests and diseases of rice plant can spread quickly and attack at all area of agriculture land. Since the number of experts in the pests and diseases of rice plant is limited, especially in a remote area, expert system then can be a smart solution for replacing the extensionist to decide what kind of pests or diseases that have attacked the rice plant. This paper presents a design and implementation of an expert system based on web application for diagnozing pests and diseases of the rice plant so that support system then still can be performed to provide the farmers with a correct decision. The knowledge representation model in this study used production rule and forward chaining based on symptoms or characteristics from attacked rice plant. The certainty factors method was used to define the expert confidence level for each symptom. This expert system testing was done by 15 person of non-extensionist of agriculture and 20 person of agriculture extensionist for observing 12 sample of images of the infected rice plant. The testing result showed that the accuracy level of this system is 73,81%. Meaning that this expert system can help farmers determining the pests or diseases of rice plant.