Integrating Autoencoders with Recursive Elastic Tree for Enhanced Fetal Health Monitoring: A Feature Engineering and Classification Approach

Abstract

This paper establishes an assimilated Autoencoder with a Recursive Elastic Tree framework that improves feature engineering and classification for monitoring fetal health during pregnancy. Autoencoders, a deep learning technique, can be utilized to find an effective and non-linear feature representation for high-dimensional fetal health data and detect complex patterns, that would otherwise have gone undetected with traditional methods. After dimensionality reduction, feature transformation, and relevance detection, RET selects the appropriate features for classification that bring about improved accuracy and increased model robustness. Putting together Autoencoders with RET resolves the setbacks of data noise, overfitting, and computational inefficiencies, thus providing a rich tool for early risk assessments and predictions of fetal health problems. The new tactic provides an essential scalable and accurate solution when dealing with large and complex high-dimensional datasets in general healthcare applications.