Maternal Mortality Rate (MMR) in Indonesia Intercensal Population Survey (SUPAS) was considered high (2015). To detect the pregnancy risk, the Public Health Center ( Puskesmas ) applies a Poedji Rochjati… Click to show full abstract
Maternal Mortality Rate (MMR) in Indonesia Intercensal Population Survey (SUPAS) was considered high (2015). To detect the pregnancy risk, the Public Health Center ( Puskesmas ) applies a Poedji Rochjati Screening Card (KSPR) demonstrating 20 features. In addition to KSPR, pregnancy risk monitoring has been assisted with a pregnancy control card consisting of 117 features. Because of the differences in the number of features between the two control cards, it is necessary to make agreements between them. To achieve this goal, there are several challenges that must be resolved, including: determining the most influential features, exploring the links among features on the KSPR and pregnancy control cards, and building a Machine Learning Model for predicting pregnancy risk. Evaluation of the most influential features is performed by using Correlation-based Feature Selection (CFS) and C5.0 algorithm, each of which produces 14 features (10 features that intersect with KSPR) and 20 features (13 features that intersect with KSPR). From the union operation in the features produced by the two techniques, there are 13 features intersecting with the attributes of KSPR. The features of operations are also utilized to form Machine Learning Model. Based on the experiment, the accuracy of the XGBoost algorithm demonstrated the greatest results of 94% followed by Random Forest, Naive Bayes, and k-Nearest Neighbor algorithms, each at 87%, 66%, and 60%. Interpretability aspects are built based on SHAP and LIME to provide insight contributing to each feature for model. The similarity feature generated in the two interpretation approaches confirmed that Cesar was dominant in determining pregnancy risk.
               
Click one of the above tabs to view related content.