Implementation of case-based reasoning and nearest neighbor similarity for peanut disease diagnosis

Abstract

This research discusses the development of an expert system to diagnose peanut disease using Case-Based Reasoning (CBR) and Nearest Neighbor Similarity. CBR is a computer reasoning system using old knowledge to overcome new problems. It provides solutions by looking at the closest old case to new case. The diagnosis process is carried out by entering a new case containing the symptoms to be diagnosed into the system, then calculating similarity values between new cases on a case base using the nearest neighbor method. The average test results of the system to make an initial diagnosis of peanut disease indicate that the system is able to correctly recognize 100% peanut disease. Accuracy calculation uses the nearest neighbor similarity method with a threshold of 0.5, 0.6 and 0.7 respectively 97.22, 88.89%, and 80.55%.