Comparison of the nearest neighbor algorithm and C4.5 for the retrieval on case-based reasoning process (case study: children respiratory disorders)

• Published in: Jurnal Mantik Vol. 8; no. 1; pp. 174 - 185
• Main Authors: Tursina, Tursina, Septiriana, Rina
• Format: Journal Article
• Language: English
• Published: 31.05.2024
• ISSN: 2685-4236; 2685-4236
• DOI: 10.35335/mantik.v8i1.5018

The diagnosis of respiratory problems was usually made through direct consultation with a pediatric respiratory specialist or by studying several previous respiratory disorders cases. These cases were gleaned from prior experiences or the knowledge of subject-matter experts. Case-Based Reasoning (CBR) is the processing of diagnosing a patient based on past cases or expertise. Retrieve, reuse, revise, and retain are some of the steps of case-based reasoning. The retrieval stage of CBR was where the classification method searches for similarity values. Numerous algorithms exist for classification techniques, such as C4.5 and the Nearest Neighbour algorithm. This study compares the similarities between the C4.5 and Nearest Neighbor algorithms. The Nearest Neighbour approach was used to search for similarity, and the results show that 99.33% of the items classified based on learning data were nearest to the object. By contrast, the accuracy value for the C4.5 approach was 100%.