Plagiarism_Radiyati_Umi_Partan_Cardiac ArrhythmiasClassificationUsing Deep Neural Networksand Principle Component AnalysisAlgorithm

Partan, Radiyati Umi (2018) Plagiarism_Radiyati_Umi_Partan_Cardiac ArrhythmiasClassificationUsing Deep Neural Networksand Principle Component AnalysisAlgorithm. Int. J. Advance Soft Compu. A.

[thumbnail of Plagiat_Radiyati_Umi_Partan_Cardiac Arrhythmias Classification Using Deep Neural Networks and Principle Component Analysis Algorithm.pdf]
Preview
Text
Plagiat_Radiyati_Umi_Partan_Cardiac Arrhythmias Classification Using Deep Neural Networks and Principle Component Analysis Algorithm.pdf

Download (6MB) | Preview

Abstract

A primary diagnostic tool for cardiovascular diseases is electrocardiogram (ECG). The ECG evaluation for predicting cardiac arrhythmiasfrom databases, resulting in a comparable or even higher accuracy than experienced examiners. In this paper,automatic classification for cardiac arrhythmiais developed forproducing higher accuracy by using combination between DNNs andPCAtechnique. The algorithm developedbased onthe following stages such as,data preparation,datareduction, feature extractionand classification of rhythms. DNNs structure utilize Soft-max regression on the top of output layer and Cross-entropy as a cost function. To validate the method on the well-known MIT-BIH arrhythmia database is used. In the experiment 18 classifierstructure with several activation functions are created, to analyze the classifier performance. To benchmark, the performance of DNNs algorithm is compared to SVM algorithm in terms of accuracy. The result obtained show that the proposed method provides good accuracyagainstto MLPand SVM, about 97.7%, 95.56% and 79.51% respectivelywith less expert interaction.

Item Type: Other
Subjects: #3 Repository of Lecturer Academic Credit Systems (TPAK) > Results of Ithenticate Plagiarism and Similarity Checker
Divisions: 04-Faculty of Medicine > 11702-Internal Medicine (Sp
Depositing User: dr radiyati umi partan
Date Deposited: 27 Dec 2019 08:30
Last Modified: 27 Dec 2019 08:30
URI: http://repository.unsri.ac.id/id/eprint/19759

Actions (login required)

View Item View Item