Machine Learning Algorithm and GESTOSIS Score Assisted High Risk Pregnancy Induced Hypertension Prediction

Abstract

This GESTOSIS scale is a validated tool which aids in predicting high risk women for Pregnancy-Induced Hypertension (PIH). This study aimed to use machine learning algorithms to determine the efficacy of the GESTOSIS score in predicting PIH. A prospective observational study was conducted on 70 pregnant women. The features in GESTOSIS scale classified as mild, moderate, or severe. The results showed that the Adaptive Boosting (AB) model precisely predicts the PIH with an accuracy range between 97% to 99% based on the results of regression and classification prediction models respectively with a true -positive rate (TPR) of 90%. The GESTOSIS score is concluded to be a simple scale that can be administered by all front-line health workers in the community without intrusive procedures.