Diagnosis System of Toddler Diseases Using Forward Chaining and Case-Based Reasoning

Abstract

Toddlers are children aged 12-36 months. This study aims to diagnose toddler diseases using forward chaining and Case-Based Reasoning (CBR). There are 16 types of toddler diseases. This study consist of two steps, i.e., diagnosis using forward chaining and diagnosis using CBR. Diagnosis using forward chaining generated 18 rules. These rules were used to determine toddler diseases type, and diagnosis using CBR focused on three types of CBR calculations, i.e., Nearest Neighbour Similarity (NNS), Minkowsky Distance Similarity (MDS), and Euclidean Distance Similarity (EDS). The results of system testing using 600 data, the accuracy of diagnosis were 82% and 90%, using forward chaining and CBR respectively. Based on these results, diagnosis using CBR was better than forward chaining, because CBR justified the data that were considered wrong, to be repaired by an expert and then made them as new cases.