Optimizing Fetal Heartbeat Detection with Feature Selection in Medical Data Using Machine Learning Algorithms

Read full text Summarise with VIRAT
Authors: Yuli Wahyuni, Hadiyanto -, Ridwan Sanjaya, Nendar Herdianto
Cited by: 0
References: 30

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

The interpretation of cardiotocography (CTG) for fetal health monitoring is often hampered by subjective assessments and inter-observer variability, which highlights the need for more objective computational techniques.This study aims to determine whether feature selection can improve the classification of prenatal health conditions using machine learning techniques.Using a publicly available CTG dataset with 2,126 recordings, Analysis of Variance (ANOVA) F-test and Recursive Feature Elimination (RFE) were performed to reduce the number of variables from 21 to 10.The models trained using the full feature set and the reduced feature set include Random Forest, Logistic Regression, and Support Vector Machine (SVM).The evaluation metrics used include F1-score, recall, accuracy, and precision.After feature selection, the Random Forest model showed the most significant improvement, with accuracy increasing from 92.72% to 93.19%, the macro F1-score rising from 86% to 87%, and the micro F1-score also demonstrating improved performance.Although the magnitude of this improvement is relatively small, these findings suggest that feature reduction can effectively reduce computational complexity while providing incremental improvements in model performance.In addition, balanced classification between Normal and Pathological classes is maintained with the smaller feature set.Overall, the results indicate that a focused feature selection approach can improve the effectiveness of fetal health categorization models without compromising prediction quality; however, further research is needed to validate the clinical use of these features.

Fields of Study & Subjects

Artificial Intelligence in Healthcare ECG Monitoring and Analysis Heart Rate Variability and Autonomic Control

Publication details

Journal / source
Mathematical Modelling and Engineering Problems
Volume
13
Issue
2
Pages
371–380
Publication date
15 March 2026
Language
EN