PENINGKATAN KINERJA KLASIFIKASI ARITMIA BERBASIS DEEP FEATURES DAN TEACHING-LEARNING BASED OPTIMIZATION

SARI, RIRIN PURNAMA and Nurmaini, Siti (2025) PENINGKATAN KINERJA KLASIFIKASI ARITMIA BERBASIS DEEP FEATURES DAN TEACHING-LEARNING BASED OPTIMIZATION. Undergraduate thesis, Sriwijaya University.

Preview	Image RAMA_56201_09011282126096_cover.jpg - Cover Image Available under License Creative Commons Public Domain Dedication. Download (499kB) | Preview
	Text RAMA_56201_09011282126096.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (4MB) | Request a copy
	Text RAMA_56201_09011282126096_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (8MB) | Request a copy
	Text RAMA_56201_09011282126096_0002085908_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (572kB)
	Text RAMA_56201_09011282126096_0002085908_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (714kB) | Request a copy
	Text RAMA_56201_09011282126096_0002085908_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (323kB) | Request a copy
	Text RAMA_56201_09011282126096_0002085908_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (2MB) | Request a copy
	Text RAMA_56201_09011282126096_0002085908_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (122kB) | Request a copy
	Text RAMA_56201_09011282126096_0002085908_06_ref.pdf - Bibliography Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (206kB) | Request a copy
	Text RAMA_56201_09011282126096_0002085908_07_lamp.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (903kB) | Request a copy

Abstract

This study developed an arrhythmia classification model based on electrocardiogram (ECG) signals using time-domain RR-interval feature extraction techniques and the implementation of Teaching-Learning-Based Optimization (TLO) for feature selection. The datasets employed include the MIT-BIH Arrhythmia Database, MIT-BIH Normal Sinus Rhythm Database, MIT-BIH Atrial Fibrillation Database, Lobachevsky University Database, and QT Database. The research process began with signal pre-processing using discrete wavelet transform (DWT) for denoising, R-peak detection, and feature extraction of RR-interval parameters, namely mean_RR, std_RR, RMS_RR, NN50, pNN50, cv_RR, min_RR, and max_RR. TLO was employed for feature selection to identify the most relevant subset of features for classification. The models tested included CNN, DNN, LSTM, and BiLSTM, each evaluated under two scenarios: with and without TLO-based feature selection. Model performance was evaluated using metrics of accuracy, recall, precision, and specificity. The findings demonstrate that the application of TLO enhances the performance of most models. The DNN+TLO model achieved the highest test accuracy of 78%, followed by CNN+TLO and LSTM+TLO, both achieving 75%. Notable improvements were also observed in specificity, with DNN+TLO and CNN+TLO attaining 94% and 93%, respectively

Item Type:	Thesis (Undergraduate)
Uncontrolled Keywords:	Klasifikasi aritmia, Elektrokardiogram, Time-domain RR-interval, Teaching-learning based optimization
Subjects:	Q Science > Q Science (General) > Q334-342 Computer science. Artificial intelligence. Algorithms. Robotics. Automation.
Divisions:	09-Faculty of Computer Science > 56201-Computer Systems (S1)
Depositing User:	Ririn Purnama Sari
Date Deposited:	16 Jul 2025 01:46
Last Modified:	16 Jul 2025 01:46
URI:	http://repository.unsri.ac.id/id/eprint/178640

Actions (login required)

View Item