PENINGKATAN KINERJA DELINEASI SINYAL ELEKTRODIAGRAM SINGLE-LEAD UNTUK MENDETEKSI ABNORMAL RHYTHM BERBASIS DEEP LEARNING

PUTRA, MUHAMMAD HAFIZH PERMANA and Nurmaini, Siti (2022) PENINGKATAN KINERJA DELINEASI SINYAL ELEKTRODIAGRAM SINGLE-LEAD UNTUK MENDETEKSI ABNORMAL RHYTHM BERBASIS DEEP LEARNING. Undergraduate thesis, Sriwijaya University.

	Text RAMA_56201_09011281924027.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (5MB) | Request a copy
	Text RAMA_56201_09011281924027_TURNITIN.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (19MB) | Request a copy
Preview	Text RAMA_56201_09011281924027_0002085908_01_front_ref.pdf - Accepted Version Available under License Creative Commons Public Domain Dedication. Download (581kB) | Preview
	Text RAMA_56201_09011281924027_0002085908_02.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (235kB) | Request a copy
	Text RAMA_56201_09011281924027_0002085908_03.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (326kB) | Request a copy
	Text RAMA_56201_09011281924027_0002085908_04.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (4MB) | Request a copy
	Text RAMA_56201_09011281924027_0002085908_05.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (85kB) | Request a copy
	Text RAMA_56201_09011281924027_0002085908_06_ref.pdf - Bibliography Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (114kB) | Request a copy
	Text RAMA_56201_09011281924027_0002085908_07_lamp.pdf - Accepted Version Restricted to Repository staff only Available under License Creative Commons Public Domain Dedication. Download (429kB) | Request a copy

Abstract

Electrodiagram (ECG) is a graph that depicts a recording of the electrical activity of the heart through electrodes placed on the body. EKG is one of the medical devices that has a function in assisting the diagnosis, especially in diseases related to the heart. Cardiac conditions can be detected by pattern analysis of the morphology of the ECG signal, both normal and abnormal conditions. In the medical world, the state of the heart can be identified by analyzing the changing patterns of heartbeats or rhythms consisting of P waves, Complex QRS and T. The method used in this study uses guidance between the Convolutional Neural Network (CNN) as a feature extractor and the Recurrent Neural Network. (RNN) Long Short Term Memory (LSTM) architecture as a wave classifier. There are two types of LSTM used in this study as a comparison, namely unidirectional and bidirectional (BiLSTM) with parameter values optimized using the grid search algorithm including epoch, batch size, and learning rate resulting in a total of 36 models. The CNN and Bi-LSTM models produce the highest evaluation in the 5-wave class scenario with a Recall value of 99.92%, Precision 99.93%, Specificity 99.98%, Accuracy 99.97%, F1 score 99.92%. Apart from that, this research also classifies abnormal waves namely, atrial fibrillation and atrial flutter from the Lobachevsky University Database (LUDB) data using the best BiLSTM model that has been built and the existing rule definitions.

Item Type:	Thesis (Undergraduate)
Uncontrolled Keywords:	KINERJA, DELINEASI, SINYAL, ELEKTRODIAGRAM
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:	Mr Muhammad Hafizh Permana Putra
Date Deposited:	09 Jan 2023 02:17
Last Modified:	09 Jan 2023 02:17
URI:	http://repository.unsri.ac.id/id/eprint/85158

Actions (login required)

View Item