Fetal Heart Rate Classification and Comparative Analysis Using Cardiotocography Data and Known Classifiers

Abstract

The problem of fetal distress usually become one of the major reason of complication during child delivery. Fetal heart rate (FHR) is one of the pivotal ways to identify the occurrence of fetal distress. Cardiotocography (CTG) is the most widely practiced technique to record FHR. Improper analysis of CTG's graph may lead to serious loss. This study presents six classification algorithms: Decision Tree (DT), K-Nearest Neighbors (KNN), Logistic Regression (LR), Support Vector Machine (SVM), Random Forest (RF) and Naïve Bayes (NB), used for the classification of CTG data. To improve the performance of the classifiers, a co-relation based feature selection technique is employed over the dataset to remove the unnecessary attributes. The performance of the classification algorithms is evaluated using evaluation metrics: Accuracy, Precision, Recall, and F-measure. The results revealed that Naïve Bayes achieved 83.06% accuracy, 92.20% precision, 83.10% recall and 86.90% f-measure.