Skripsi

DELINEASI GELOMBANG P DAN RR-INTERVAL PADA SINYAL ELEKTROKARDIOGRAM SINGLE-LEAD MENGGUNAKAN METODE RECURRENT NEURAL NETWORKS

---

Penilaian

---

This research aims to perform the delineation of P-waves and RR-intervals in single-lead electrocardiogram (ECG) signals using Recurrent Neural Networks (RNN). The ECG signals consist of P, Q, R, S, T, and U waves. The delineation process is conducted to identify waves in the EKG signal, dividing the data into eight classes, namely Pwave, Poff-Qon, Qon-Rpeak, Rpeak-Qoff, Qoff-Ton, Twave, Toff-Pon, and Zeropad. The utilization of deep learning methods in delineation aims to reduce interpretation errors. In this study, a computer-based delineation system employs a combination of CNN-BiLSTM deep learning methods. Delineation is carried out for eight wave classes, with a total of 312 designed models, each trained and tested using QTDB data. Each model is constructed with variations in hidden layer parameters, batch size, learning rate, and epoch to achieve optimal results. The delineation process of medical image signals in EKG with the CNN-BiLSTM architecture shows the best results in trials using a CNN with 7 layers and 1 layer of BiLSTM. The fourth model in this architecture exhibits a sensitivity of 93.59%, precision of 94.94%, specificity of 99.52%, accuracy of 99.15%, error rate of 0.85%, and an F1-Score of 94.20%

---

Availability

Inventory Code	Barcode	Call Number	Location	Status
2407000877	T138216	T1382162024	Central Library (Referens)	Available but not for loan - Not for Loan

Detail Information

Series Title

-

Call Number

T1382162024

Publisher

Inderalaya : Prodi Sistem Komputer, Fakultas Ilmu Komputer Uniersitas Sriwijaya., 2023

Collation

xiv, 68 hlm.; Ilus.; 29 cm

Language

Indonesia

ISBN/ISSN

-

Classification

004.650 7

Content Type

Text

Media Type

unmediated

Carrier Type

-

Edition

-

Subject(s)

Prodi Sistem Komputer
Jaringan komunikasi

Specific Detail Info

-

Statement of Responsibility

SEPTA

Other version/related

File Attachment

Comments

---

You must be logged in to post a comment