Penerapan Algoritma Machine Learning Random Forest untuk Prediksi Risiko Konversi Sindrom Terisolasi Klinis Menjadi Multiple Sclerosis
- Riki Ripaldi Universitas Bina Sarana Informatika, Bekasi, Indonesia
- Leonardo Sebastian Tambunan Universitas Bina Sarana Informatika, Bekasi, Indonesia
-
Samuel Edowardo
*
Universitas Bina Sarana Informatika, Bekasi,
Indonesia
- Syifa Nur Rahkmah Universitas Bina Sarana Informatika, Bekasi, Indonesia
- Imam Sutoyo Universitas Bina Sarana Informatika, Bekasi, Indonesia
- Findi Ayu Sariasih Universitas Bina Sarana Informatika, Bekasi, Indonesia (*) Corresponding Author
Abstract
Clinically Isolated Syndrome (CIS) is an initial neurological episode potentially developing into Multiple Sclerosis (MS), a chronic neurodegenerative disorder of the central nervous system. Early detection of risk factors for CIS to MS conversion is crucial for supporting timely medical interventions and slowing down disease progression. This study aims to develop a risk prediction model for CIS to MS conversion using a Machine Learning algorithm, comprehensively evaluate the model's performance, and implement it as a web-based clinical decision support system. The research employs a machine learning approach utilizing the Random Forest Classifier to predict the conversion risk using the public dataset Conversion Predictors of CIS to Multiple Sclerosis. The dataset comprises 273 patients with clinical variables including demographics, initial symptom characteristics, Magnetic Resonance Imaging (MRI) findings across various brain regions and the spinal cord, and Oligoclonal Bands (OCB) test results. The methodology involved addressing class imbalance using weight adjustments, cross-validation, and implementing a custom threshold of 0.57 to minimize false positives, ensuring clinical diagnostic safety. Test results demonstrate that the Random Forest model achieved optimal performance with an Accuracy of 81.82%, an F1-Score of 0.82, and an Area Under the Curve (AUC) of 0.9140, indicating excellent discriminative capability. Feature Importance analysis revealed that Oligoclonal Bands (OCB), Initial Symptoms (specifically sensory and visual disturbances), and MRI lesions (especially Periventricular) are the most influential predictors. The model is subsequently implemented into a web-based prediction system to facilitate interactive risk assessment by medical professionals. This implementation serves as an accurate and explainable prototype of a Clinical Decision Support System.
References
A. Nugroho and R. Rahmawati, “Implementasi Machine Learning untuk deteksi dini penyakit neurologis,” Jurnal MISI, vol. 7, no. 2, pp. 33–40, 2023.
A. H. Wibowo, “Analisis performa model pembelajaran mesin dalam prediksi penyakit berdasarkan data medis,” Jurnal Sistem Informasi dan Komputerisasi Akuntansi (JUSIKA), vol. 6, no. 1, pp. 23–32, 2022.
C. Zhang and others, “Data imbalance handling in medical classification using Random Forest ensemble,” Appl Soft Comput, vol. 122, p. 108919, 2022.
D. Gebre, “Conversion Predictors of CIS to Multiple Sclerosis,” 2022, Kaggle Dataset.
E. Marini and others, “Feature importance in machine learning models for neurological diseases,” Artif Intell Med, vol. 131, p. 102365, 2022.
F. M. Collins and R. H. Morris, “Clinical implications of early CIS conversion: A data-driven analysis,” Neurol Res Int, vol. 2023, pp. 1–10, 2023.
H. J. Kim, “Optimization of Random Forest hyperparameters for medical data classification,” Comput Biol Med, vol. 140, p. 105086, 2022.
I. P. Wirawan and A. Setiawan, “Penerapan algoritma Random Forest untuk prediksi penyakit stroke,” Jurnal Teknologi Informasi dan Komputer, vol. 9, no. 2, pp. 145–153, 2023.
B. Rahmani and T. Yu, “Risk Stratification of Clinically Isolated Syndrome for Multiple Sclerosis Development,” Med Res Arch, vol. 13, no. 10, 2025.
S. Mendanha and K. R. Gudibandi, “Comparative Analysis of Machine Learning Algorithms for Conversion Predictors of Clinically Isolated Syndrome (CIS) to Multiple Sclerosis (MS),” in IEEE Conference on Data Analysis and Deep Learning, 2025.
N. Puspitasari and D. Riyadi, “Integrasi sistem prediksi penyakit berbasis web dengan algoritma Random Forest,” Jurnal Manajemen Informatika dan Sistem Informasi (MISI), vol. 8, no. 1, pp. 45–54, 2024.
H. Hu, L. Ye, P. Wu, Z. Shi, G. Chen, and Y. Li, “Exploring factors driving the evolution of chronic lesions in multiple sclerosis using machine learning,” Eur Radiol, 2025.
Z. Wang and others, “Medical data-driven prediction of multiple sclerosis using ensemble learning,” Sci Data, vol. 9, no. 221, pp. 1–10, 2022.
S. Sima and others, “Predictive modeling for conversion from clinically isolated syndrome to multiple sclerosis using random forest,” Neuroinformatics Journal, vol. 19, no. 4, pp. 511–523, 2023.
Y. Zhao and others, “Random Forest-based clinical decision support for multiple sclerosis,” Front Neurosci, vol. 16, pp. 221–234, 2022.
S. Al-Khayer and others, “Deep learning and Random Forests for early detection of MS lesions in MRI,” Comput Biol Med, vol. 144, p. 105325, 2022.
M. Goyal, A. Singh, and P. Kumar, “Machine learning-based early diagnosis of multiple sclerosis using MRI data,” Biocybern Biomed Eng, vol. 42, no. 3, pp. 857–871, 2022.
H. Khan, H. C. Woodruff, D. L. Giraldo, and L. Werthen-Brabants, “Leveraging Hand-Crafted Radiomics on Multicenter FLAIR MRI for Predicting Disability Progression in People with Multiple Sclerosis,” medRxiv, 2025.
S. P. Rao, “Feature engineering and model interpretability in multiple sclerosis prediction,” Int J Med Inform, vol. 157, p. 104651, 2022.
Z. Kaur and others, “Explainable machine learning for MS progression analysis,” Front Artif Intell, vol. 5, p. 902315, 2023.
Bila bermanfaat silahkan share artikel ini
Berikan Komentar Anda terhadap artikel Penerapan Algoritma Machine Learning Random Forest untuk Prediksi Risiko Konversi Sindrom Terisolasi Klinis Menjadi Multiple Sclerosis
Pages: 23 - 32
Copyright (c) 2025 Riki Ripaldi, Leonardo Sebastian Tambunan, Samuel Edowardo, Syifa Nur Rahkmah, Imam Sutoyo, Findi Ayu Sariasih
This work is licensed under a Creative Commons Attribution 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under Creative Commons Attribution 4.0 International License that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (Refer to The Effect of Open Access).
