Decision Tree Model for Maternal Risk Classification Based on Kartu Skor Poedji Rochjati

Dyah Megawati Surip Solekhah (Universitas Gadjah Mada)
Annisa Maulida Ningtyas (Universitas Gadjah Mada)

JUITA : Jurnal Informatika
JUITA Vol. 14 Issue 2, July 2026

Maternal mortality remains a serious public health issue. This can be prevented by taking preventive measures through early identification of pregnancy risks. This study aims to develop a classification model for screening maternal pregnancy risks using the “Kartu Skor Poedji Rochjati” (KSPR) as a clinical basis for data labelling. This research applies the Cross-Industry Standard Process for Medical Data Mining (CRISP-MED-DM) methodology to ensure a systematic and clinically relevant modelling process. A dataset containing 998 medical records from the Maternal Health and High-Risk Pregnancy Dataset was used, with rule-based labelling adapted from KSPR to classify maternal risk into three categories: Low-Risk Pregnancy, High-Risk Pregnancy, and Very High-Risk Pregnancy. Experiments were conducted with three Decision Tree models, namely the Baseline Model, Model with SMOTE, and Model with Class Weighting. Based on these experiments, it was found that the Decision Tree algorithm enhanced with the Synthetic Minority Oversampling Technique (SMOTE) to overcome class imbalance was the most optimal model. This model achieved balanced performance across all classes with an accuracy of 0.86 and a weighted average F1-Score of 0.87.