Expert System for Diagnosing Bacterial Infections in Children Using the Dempster Shafer Method

Naufal Yeza Ananda Yeza (Malikussaleh University)
Mukti Qamal (3Department of Informatics, Malikussaleh University, Bukit Indah, Lhokseumawe, 24353, Indonesia,)
Zahratul Fitri (Informatics Engineering, Malikussaleh University, Lhokseumawe, 24353, Indonesia)
Ellya Nova Lubis (Department of Pediatrics, Mitra Sejati Hospital, Medan, Indonesia)

Proceedings of International Conference on Multidisciplinary Engineering (ICOMDEN)
Vol. 2 (2024): Proceedings of International Conference on Multidisciplinary Engineering (ICOMDEN)

This study seeks to apply the Dempster-Shafer Method within an expert system designed for diagnosing bacterial infections in children. Developed as an Android application, the system not only aids in identifying the type of bacterial infection based on observed symptoms but also improves efficiency regarding time and costs while enhancing parental awareness of various bacteria that could impact their children's health. Symptoms evaluated include fever, assigned a weight of 1, diarrhea (from mild to bloody) with a weight of 0.8, nausea and vomiting with a weight of 1, headaches with a weight of 0.8, and severe abdominal pain also weighted at 0.8. The Dempster-Shafer Method processes uncertain information through calculations based on symptom weights defined by pediatric health specialists. Results reveal that this method effectively supports the diagnosis of bacterial infections in children, yielding the highest probability values for Escherichia coli (P02) and Salmonella typhi (P03), calculated at 1.33 and subsequently rounded to 1.0, corresponding to 100%, thus offering a more precise and advantageous approach for medical decision-making.