MARSANTI, TIKA PUTRI (2026) ANALISIS FAKTOR RISIKO STUNTING MENGGUNAKAN LOGISTIC REGRESSION BERBASIS ODDS RATIO DAN SUPPORT VEKTOR MACHINE. S1 thesis, Universitas Mercu Buana Jakarta.
|
Text (HAL COVER)
Cover.pdf Download (905kB) | Preview |
|
![[img]](https://repository.mercubuana.ac.id/style/images/fileicons/text.png) |
Text (BAB I)
BAB I.pdf Restricted to Registered users only Download (176kB) |
|
|
Text (BAB II)
BAB II.pdf Restricted to Registered users only Download (439kB) |
|
|
Text (BAB III)
BAB III.pdf Restricted to Registered users only Download (284kB) |
|
|
Text (BAB IV)
BAB IV.pdf Restricted to Registered users only Download (531kB) |
|
|
Text (BAB V)
BAB V.pdf Restricted to Registered users only Download (159kB) |
|
|
Text (DAFTAR PUSTAKA)
DAFTAR PUSTAKA.pdf Restricted to Registered users only Download (213kB) |
|
|
Text (LAMPIRAN)
LAMPIRAN.pdf Restricted to Registered users only Download (467kB) |
Abstract
Stunting remains a major public health concern in Indonesia. Early identification of stunting risk at the primary healthcare level is still challenging due to the limited utilization of anthropometric data and the need for predictive methods that provide both high accuracy and clinical interpretability. This study aims to compare the performance of Logistic Regression, Logistic Regression with Polynomial Features, and Support Vector Machine (SVM) in predicting stunting risk among children under five years of age. In addition, the study identifies stunting risk factors through Odds Ratio analysis based on the Logistic Regression model. The study employed secondary anthropometric data of children aged 0–60 months, including gender, age, body weight, body length/height, birth weight, birth length, weight gain status, and two engineered features, namely Body Mass Index (BMI) and Birth Ratio. The research process included data preprocessing, class imbalance handling using SMOTE, model development, threshold tuning, Odds Ratio estimation with a 95% confidence interval, and evaluation using Accuracy, Precision, Recall, F1-Score, ROC-AUC, and PR-AUC. Model validation was conducted using Stratified 5-Fold Cross-Validation. The results indicate that Logistic Regression with Polynomial Features achieved the best performance, with an F1-Score of 0.95, ROC-AUC of 0.99, and PRAUC approaching 1.00. Odds Ratio analysis identified body height, weight gain status, and birth length as significant risk factors for stunting, whereas birth weight and gender were not significantly associated. The proposed dual-model approach combines a prediction model with an interpretable Logistic Regression model for clinical risk assessment. The findings are expected to support early identification of stunting risk in primary healthcare settings. Keywords: stunting, Logistic Regression, Polynomial Features, Support Vector Machine, Odds Ratio. Stunting masih menjadi salah satu permasalahan kesehatan masyarakat di Indonesia. Deteksi dini risiko stunting di pelayanan kesehatan primer belum optimal karena pemanfaatan data antropometri masih terbatas serta diperlukan metode prediksi yang memiliki akurasi tinggi dan mudah diinterpretasikan. Penelitian ini bertujuan membandingkan kinerja Logistic Regression, Logistic Regression dengan Polynomial Features, dan Support Vector Machine (SVM) dalam memprediksi risiko stunting pada balita serta mengidentifikasi faktor risiko melalui analisis Odds Ratio pada Logistic Regression. Penelitian menggunakan data sekunder antropometri balita usia 0–60 bulan yang mencakup jenis kelamin, usia, berat badan, tinggi atau panjang badan, berat badan lahir, panjang badan lahir, status kenaikan berat badan, serta dua fitur hasil rekayasa, yaitu Indeks Massa Tubuh (IMT) dan Rasio Lahir. Tahapan penelitian meliputi preprocessing data, penanganan ketidakseimbangan kelas menggunakan SMOTE, pengembangan model, threshold tuning, perhitungan Odds Ratio beserta interval confidence 95%, serta evaluasi menggunakan Accuracy, Precision, Recall, F1-Score, ROC-AUC, dan PR-AUC. Validasi dilakukan dengan Stratified 5-Fold Cross Validation. Hasil penelitian menunjukkan bahwa Logistic Regression dengan Polynomial Features memberikan performa terbaik dengan F1-Score sebesar 0,95, ROC-AUC sebesar 0,99, dan PR-AUC mendekati 1,00. Analisis Odds Ratio menunjukkan bahwa tinggi badan, status kenaikan berat badan, dan panjang badan lahir berpengaruh signifikan terhadap risiko stunting, sedangkan berat badan lahir dan jenis kelamin tidak signifikan. Penelitian ini menerapkan dua model, yaitu model prediksi untuk memperoleh performa klasifikasi terbaik dan model interpretasi berbasis Logistic Regression untuk menghasilkan Odds Ratio yang mudah dipahami. Hasil penelitian diharapkan dapat mendukung identifikasi dini risiko stunting pada pelayanan kesehatan primer. Kata Kunci: stunting, Logistic Regression, Polynomial Features, Support Vector Machine, Odds Ratio.
Actions (login required)
 |
View Item |
