Classification of normal and malignant ventricular arrhythmia ECG rhythms using machine learning tools
The increasing prevalence of heart disease among individuals is a call for alarm, especially since heart disease remains a leading cause of death worldwide. As such, it is of utmost importance to identify any irregularity in the functioning of the heart, at the earliest. Arrhythmia is one such irreg...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project (FYP) |
| Language: | English |
| Published: |
Nanyang Technological University
2024
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/10356/175037 |
| _version_ | 1824454384272015360 |
|---|---|
| author | Prabhakaran, Sahithya |
| author2 | Vidya Sudarshan |
| author_facet | Vidya Sudarshan Prabhakaran, Sahithya |
| author_sort | Prabhakaran, Sahithya |
| collection | NTU |
| description | The increasing prevalence of heart disease among individuals is a call for alarm, especially since heart disease remains a leading cause of death worldwide. As such, it is of utmost importance to identify any irregularity in the functioning of the heart, at the earliest. Arrhythmia is one such irregularity in the functioning of the heart. While medical professionals visually inspect Electrocardiogram (ECG) readings to diagnose the presence of arrhythmia, its time-consuming nature and the possibility of human error, motivated various research to be done on the use of machine learning to facilitate the classification of arrhythmia. This study focuses on the implementation and comparison of models established in published research papers. The deep residual Convolutional Neural Network (CNN) achieved the highest accuracy of 0.967 as compared to the deep residual CNN with transfer learning and VFPred, which is a fusion of signal processing and Support Vector Machine (SVM). Despite its high performance in accuracy, it fell short in terms of F1 score and recall, with VFPred achieving the highest F1 score of 0.877. Hence, while the deep residual CNN demonstrated high accuracy, VFPred emerged as the favorable choice among the three, as it attained a balanced performance across the evaluation metrics. |
| first_indexed | 2025-02-19T03:21:27Z |
| format | Final Year Project (FYP) |
| id | ntu-10356/175037 |
| institution | Nanyang Technological University |
| language | English |
| last_indexed | 2025-02-19T03:21:27Z |
| publishDate | 2024 |
| publisher | Nanyang Technological University |
| record_format | dspace |
| spelling | ntu-10356/1750372024-04-19T15:45:23Z Classification of normal and malignant ventricular arrhythmia ECG rhythms using machine learning tools Prabhakaran, Sahithya Vidya Sudarshan School of Computer Science and Engineering vidya.sudarshan@ntu.edu.sg Computer and Information Science Machine learning The increasing prevalence of heart disease among individuals is a call for alarm, especially since heart disease remains a leading cause of death worldwide. As such, it is of utmost importance to identify any irregularity in the functioning of the heart, at the earliest. Arrhythmia is one such irregularity in the functioning of the heart. While medical professionals visually inspect Electrocardiogram (ECG) readings to diagnose the presence of arrhythmia, its time-consuming nature and the possibility of human error, motivated various research to be done on the use of machine learning to facilitate the classification of arrhythmia. This study focuses on the implementation and comparison of models established in published research papers. The deep residual Convolutional Neural Network (CNN) achieved the highest accuracy of 0.967 as compared to the deep residual CNN with transfer learning and VFPred, which is a fusion of signal processing and Support Vector Machine (SVM). Despite its high performance in accuracy, it fell short in terms of F1 score and recall, with VFPred achieving the highest F1 score of 0.877. Hence, while the deep residual CNN demonstrated high accuracy, VFPred emerged as the favorable choice among the three, as it attained a balanced performance across the evaluation metrics. Bachelor's degree 2024-04-18T23:43:17Z 2024-04-18T23:43:17Z 2024 Final Year Project (FYP) Prabhakaran, S. (2024). Classification of normal and malignant ventricular arrhythmia ECG rhythms using machine learning tools. Final Year Project (FYP), Nanyang Technological University, Singapore. https://hdl.handle.net/10356/175037 https://hdl.handle.net/10356/175037 en SCSE23-0716 application/pdf Nanyang Technological University |
| spellingShingle | Computer and Information Science Machine learning Prabhakaran, Sahithya Classification of normal and malignant ventricular arrhythmia ECG rhythms using machine learning tools |
| title | Classification of normal and malignant ventricular arrhythmia ECG rhythms using machine learning tools |
| title_full | Classification of normal and malignant ventricular arrhythmia ECG rhythms using machine learning tools |
| title_fullStr | Classification of normal and malignant ventricular arrhythmia ECG rhythms using machine learning tools |
| title_full_unstemmed | Classification of normal and malignant ventricular arrhythmia ECG rhythms using machine learning tools |
| title_short | Classification of normal and malignant ventricular arrhythmia ECG rhythms using machine learning tools |
| title_sort | classification of normal and malignant ventricular arrhythmia ecg rhythms using machine learning tools |
| topic | Computer and Information Science Machine learning |
| url | https://hdl.handle.net/10356/175037 |
| work_keys_str_mv | AT prabhakaransahithya classificationofnormalandmalignantventriculararrhythmiaecgrhythmsusingmachinelearningtools |