Desain Dempster Shafer dan Fuzzy Expert System dalam Mendeteksi Dini Penyakit Stroke

• Main Author: laurentinus laurentinus
• Format: Article
• Language: Indonesian
• Published: Ikatan Ahli Indormatika Indonesia 2020-10-01
• Series: Jurnal RESTI (Rekayasa Sistem dan Teknologi Informasi)
• Subjects: expert system, stroke, fuzzy mamdani, dempster shafer, android
• Online Access: http://jurnal.iaii.or.id/index.php/RESTI/article/view/2227

The increasing population in Indonesia, which is 265 million people in 2018, causes an increase in the community's disease sufferers. Unfortunately, the number of hospitals in the area has not increased even though the population continues to grow, which impacts the community's lack of information and knowledge in dealing with some serious diseases such as stroke that attacks quickly. Stroke is the leading cause of disability and the number two cause of death in the world where 6.2 million people died in 2015 and is a complex medical problem that requires the diagnosis of a neurologist or internist. Still, not all doctors are in the district and provide services with fast. Temporary stroke symptoms are called transient ischemic attacks (TIA), which are warning signs before having a stroke, it requires how to recognize the signs of a stroke early and treat it as a medical emergency. Based on this problem, it is needed an expert system design that can diagnose stroke early and provide information about stroke to the community based on expert sources with an android mobile phone, making it accessible to the broader community, including in the district. The system design uses the Dempster Shafer Method to measure the uncertainty of 20 stroke symptoms. The disease slices outcome will produce a percentage of the likelihood of stroke, hypertension / high blood pressure, fever, and heart disease. As well as Fuzzy Logic as logical logic in processing 9 patient's medical history. The authors combined the two methods in providing a stroke diagnosis based on symptoms and patient history and then evaluated using several metrics, including accuracy, precision, sensitivity (recall), F-measure (F1 score), and specificity so that an expert system score was obtained of 0.786 which shows good expert system performance.