Perinatal hypoxia diagnostic system by using scalable machine learning algorithms

Abstract

Collaborating big data and machine learning approaches in healthcare can help in improving clinical decision making and treatment by identifying and accumulating accurate features. Prenatal hypoxia can also be identified by cardiotocography (CTG) monitoring that helps in identifying the condition of the fetus. Imposing the data over distributed approaches can help in fast computation to rate the fetal and mother wellbeing before delivery. Our research aims to propose and implement a scalable Machine learning Algorithm based perinatal Hypoxia diagnostic system for larger datasets. This system was implemented on the CTG dataset using python and pyspark models like SVM, Random Forest, and Logistic regression. In the proposed method experiment results contributing to spark RF are more accurate than other techniques and achieved the precision of 0.97, recall of 0.99, f-1 score of 0. 98, AUC of 0.97 and gained 97% accuracy.