Prediction of Pregnancy-Induced Hypertension Levels Using Machine Learning Algorithms

TLDR: This work shows that the accuracy achieved by the use of decision tree (90%) is better than that of support vector machine (86.667%) and logistics regression (83.334%) algorithms used in earlier work.

Abstract: Pregnancy-induced hypertension (PIH) is a foremost reason for disease and death in maternal, fetal, and neonatal babies. Women having PIH are at greater risk of intrauterine growth retardation in fetuses, premature delivery of a baby, and intrauterine death. Machine Learning has been widely used in an array of applications in the healthcare domain for analyzing data. The aim of this study by the authors is to predict the PIH levels using supervised learning algorithms with an aim to prevent PIH-related complications. The study works on a data set of about 100 pregnant women between the age group of 18–32. The data set uses 19 predictor variables like body surface area (BSA), pulse rate (PR), systolic blood pressure (SBP), and diastolic blood pressure (DBP). SBP and DBP variables are considered to predict the PIH level of the pregnant woman. This work shows that the accuracy achieved by the use of decision tree (90%) is better than that of support vector machine (86.667%) and logistics regression (83.334%) algorithms used in earlier work.