Leveraging SAMME for Improved Multi-Class Cirrhosis Diagnosis in Clinical Settings

Arum Kurnia Sulistyawati
Dyan Avando Meliala
Ajie Wibowo Soejono
Dini Sari
Marselina Endah Hiswati (Universitas Respati Yogyakarta)
Mohammad Diqi

Jurnal Informatika dan Rekayasa Perangkat Lunak
Vol. 7 No. 1 (2025): Maret

This study explores the use of the SAMME algorithm to develop a predictive model for identifying various stages of cirrhosis. The dataset includes 418 records with 20 attributes, targeting the classification of cirrhosis stages: C (censored), CL (censored due to liver transplantation), and D (death). The model achieved an overall accuracy of 94%, demonstrating high precision and recall for classes C and D. However, the precision for class CL was lower, indicating a tendency to over-predict this stage. These results validate the SAMME algorithm's potential to enhance diagnostic accuracy while highlighting the need for further refinement to address class imbalance and feature overlap. This research underscores the value of machine learning in early diagnosis and personalized treatment, suggesting future work on larger, balanced datasets and advanced feature engineering to improve model robustness and reliability in clinical applications.