Labor Prediction in pregnant woman based on ElectroMyoGram and ElectroHysteroGram using Machine Learning Techniques

The World Health organization statistics states that preterm delivery is the most important cause of infant mortality. Babies born before 37 weeks of gestational age as opposed to the usual 40 weeks are referred preterm babies. Studies suggest that at least 75% of these premies would survive if they are detected at an earlier stage and given appropriate treatment. While Preterm infants suffer chronic issues like mental retardation, dyslexia, visual impairments, and hearing impairments, Preterm labor causes placental abruption and previa, uterine overdistention, prolonged labor, and in some cases malposition. The key to cure is solely prevention and prediction. The proposed work aims at alerting pregnant women during the initial contractions of the labor using Uterine ElectroMyoGram (EMG), commonly termed as ElectroHysteroGram (EHG). The system consists of AgC12 electrodes placed above and below the naval. The signals are collected before and after 26 weeks of the gestation term, after which they are collected for 37 weeks and more than 37 weeks. After preprocessing and extracting the features, the data collected in term and preterm is classified using various machine learning classifiers such as Support Vector Machine (SVM), Naïve Bayes Classifier, K-NN Classifier, Gradient Boost, Decision Tree and it can be inferred that SVM and Gradient boost classifiers produces an accuracy rate that is better when compared to KNN, Naïve Bayes and Decision tree classifiers.