Dhivya, S and Saranya, R Enhancing Clinical Diagnostics and Patient Monitoring With Recurrent Neural Networks: A Comparative Study of LSTM and GRU Architecture. Enhancing Clinical Diagnostics and Patient Monitoring With Recurrent Neural Networks: A Comparative Study of LSTM and GRU Architectures. pp. 379-380.

[thumbnail of Dr.S.Dhivya (1).pdf] Text
Dr.S.Dhivya (1).pdf - Published Version
Download (321kB)

Abstract

Recurrent Neural Networks (RNNs), particularly Long Short-Term Memory (LSTM) and Gated Recurrent
Unit (GRU) architectures, are powerful tools for modeling clinical time-series data where sequential
information is essential. This paper explores the effectiveness of RNNs in analyzing patient vital signs,
electrocardiogram (ECG) readings, and other time-sensitive health data. The recurrent design of
these networks enables the capture of short- and long-term dependencies, vital for predicting patient
outcomes, anomaly detection, and aiding in diagnostic decisions. We review recent advancements in
RNN applications for healthcare, addressing challenges like irregular sampling, missing data, and the
need for interpretability in clinical settings. Through experiments with real-world clinical datasets, we
demonstrate that RNNs outperform traditional machine learning models, enhancing diagnostics, patient
monitoring, and personalized treatment planning. Our results highlight the significant potential of RNNs
in advancing healthcare analytics and supporting data-driven decision-making

Item Type: Article
Divisions: PSG College of Arts and Science > Department of Computer Science
Depositing User: Dr. B Sivakumar
Date Deposited: 18 Apr 2026 05:46
Last Modified: 18 Apr 2026 05:46
URI: https://ir.psgcas.ac.in/id/eprint/2806

Actions (login required)

View Item
View Item