Perbandingan Keefektifan Metode Case-Based Reasoning dan Certainty Factor dalam Sistem Pakar Diagnosis Penyakit Multiple Sclerosis

Hanifah Ekawati (STMIK Widya Cipta Dharma, Samarinda)
Ita Arfyanti (STMIK Widya Cipta Dharma, Samarinda)
Tommy Bustomi (Politeknik Negeri Samarinda, Samarinda)

JURNAL MEDIA INFORMATIKA BUDIDARMA
Vol 7, No 4 (2023): Oktober 2023

The management of complex neurological diseases such as Multiple Sclerosis (MS) requires accurate and efficient diagnostic approaches. To enhance diagnostic precision, a study has conducted a comparison between two approaches within the framework of an expert system, namely the Case-Based Reasoning (CBR) Method and the Certainty Factor (CF) Method. The primary objective of this study is to evaluate the effectiveness of these two methods in supporting the diagnosis process of Multiple Sclerosis. The Case-Based Reasoning Method is an approach that relies on past experiences to address new issues. Within an expert system, CBR utilizes knowledge from previous cases to identify diagnoses that align with the current situation. On the other hand, the Certainty Factor Method is an approach that measures the confidence level in a statement based on rules and associated confidence factors. This study makes use of a dataset containing information from previous cases related to the diagnosis of Multiple Sclerosis. By employing both of these methods, an expert system is developed to provide diagnostic recommendations based on inputted symptoms and data. The effectiveness of both approaches is evaluated through diagnostic accuracy, computational speed, and confidence levels in the generated results. Research findings indicate that both methods have their respective strengths and weaknesses. The CBR method tends to yield accurate results by referring to similar cases in the past, but it may encounter challenges in unique or rare cases. On the other hand, the Certainty Factor Method has the ability to handle uncertainty and can produce results with measurable confidence levels. However, dependence on predefined rules may limit adaptation to new cases. In conclusion, this study underscores that there is no singular perfect approach within expert systems for diagnosing Multiple Sclerosis. Both the CBR and Certainty Factor methods contribute in their own ways to improving accuracy and confidence in the diagnosis process. Therefore, integrating these two methods could be a promising direction for the development of expert systems in the future.