Arrhythmia Classification Using Hybrid and Standalone Deep Learning Models

Bhardwaj, Parkhi (2024) Arrhythmia Classification Using Hybrid and Standalone Deep Learning Models. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (2MB) | Preview
Preview	PDF (Configuration Manual) Download (1MB) | Preview

Abstract

Arrhythmias are the severe conditions which need precise and fast reports and often difficult to detect due to differences in ECG signal patterns and an imbalance between classes of data. This work: seeks to study the arrhythmias classification problem by training (i) standalone and (ii) hybrid deep learning models on the standard MIT-BIH Arrhythmia Dataset. In the study, four architectures; RNN, LSTM, GRU, and CNN connected LSTM are proposed based on the challenges which are solved in robust preprocessing techniques such as segmentation, normalization, augmentation and technique in handling class imbalance using SMOTE and class weighting. Of these, the CNN-LSTM has the highest overall performance by integrating Spatial Feature Extraction with Temporal Dependency Learning by scoring high G-mean and best accuracy, sensitivity and specificity rates of all the heartbeat classes particularly the minority ones. Comparing with the RNN and GRU models, standalone models computational efficiency has its drawbacks on long-term dependencies, therefore, the CNN-LSTM is the best option for arrhythmia detection. Hence, this research discusses the effectiveness of using automated algorithms for classification of arrhythmias and consequently the possibilities of reducing the extent to which experts’ assistance is relied upon as well as increasing the possibilities of the general availability of the data set for real-world clinical use.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Raj, Kislay UNSPECIFIED
Subjects:	Q Science > QA Mathematics > Electronic computers. Computer science T Technology > T Technology (General) > Information Technology > Electronic computers. Computer science R Medicine > Healthcare Industry H Social Sciences > HM Sociology > Information Science > Communication > Medical Informatics
Divisions:	School of Computing > Master of Science in Artificial Intelligence
Depositing User:	Ciara O'Brien
Date Deposited:	19 Jun 2025 15:29
Last Modified:	19 Jun 2025 15:29
URI:	https://norma.ncirl.ie/id/eprint/7944

Actions (login required)

View Item