Akurasi Metode Mesin Pembelajaran dalam Analisis Variabel Penting Faktor Risiko Sindrom Down

Oscar Oleta Palit; Rafi Prayoga Dhenanta; Agnes Indarwati Susanto; Adzky Matla Syawly; Atthar Luqman Ivansyah; Aditya Purwa Santika; Mochamad Ikbal Arifyanto; Fahdzi Muttaqien

doi:10.33022/ijcs.v13i5.4354

Authors

Oscar Oleta Palit Institut Teknologi Bandung https://orcid.org/0000-0003-4215-8941
Rafi Prayoga Dhenanta Institut Teknologi Bandung
Agnes Indarwati Susanto Institut Teknologi Bandung https://orcid.org/0009-0005-1053-2283
Adzky Matla Syawly Institut Teknologi Bandung https://orcid.org/0009-0008-1273-0576
Atthar Luqman Ivansyah Institut Teknologi Bandung https://orcid.org/0000-0001-9716-8982
Aditya Purwa Santika Institut Teknologi Bandung https://orcid.org/0000-0001-9525-1088
Mochamad Ikbal Arifyanto Institut Teknologi Bandung https://orcid.org/0000-0001-7205-9472
Fahdzi Muttaqien Institut Teknologi Bandung https://orcid.org/0000-0001-8970-444X

DOI:

https://doi.org/10.33022/ijcs.v13i5.4354

Keywords:

Down syndrome, case control, machine learning, accuracy, important variables

Abstract

This study aims to identify risk factors for Down syndrome using machine learning methods. Data were obtained from an epidemiological case-control study conducted at Special Needs Schools in the cities and regencies of Tangerang. Methods used include Random Forest, K-Nearest Neighbors, Support Vector Machine (SVM), Naive Bayes, K-Means, Artificial Neural Network (ANN), and Multi-Layer Perceptron (MLP). The results indicate that maternal age, paternal age, and the time interval of parents' work before the child's birth are the most influential factors in the incidence of Down syndrome. The SVM method achieved the highest accuracy of 76% with data categorized into two groups and using important variables. In addition to SVM, Naive Bayes and Random Forest methods also demonstrated good performance for analyzing epidemiological data with case-control types.