Octaviani, Vika (2017) Klasifikasi Detak Jantung Normal Dan Abnormal Berbasis Sinyal EKG Menggunakan Metode Backpropagation. Masters thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
2215205006-Master_Tesis.pdf - Accepted Version Download (2MB) | Preview |
Abstract
Kondisi fisiologis jantung manusia normal dan abnormal (arrhythmia) dapat diketahui melalui sinyal EKG (Electrocardiography). Penelitian ini bertujuan untuk melakukan pengenalan dan klasifikasi terhadap kondisi jantung manusia. Kondisi jantung tersebut yakni, normal sinus rhythm, supraventricular arrhythmia, ventricular tachyarrhythmia dan atrial fibrillation. Pemodelan sinyal menggunakan Pan-Tompkin algorithm, dimana dengan menggunakan algoritma tersebut, kita bisa mengekstrak informasi-informasi penting pada sinyal. Parameter yang didapatkan tersebut digunakan sebagai masukan bagi Backpropagation (multi layer perceptron) dengan fungsi sebagai data latih dan data uji. Data yang dipakai masing-masing subjek adalah 10 detik dengan total subjek 69 orang yang mewakili 4 kelas output. Hasil pengujian menunjukkan bahwa klasifikasi terbaik menggunakan 4 parameter inputan yaitu QRS kompleks, RR interval, nilai bpm dan Interval Q-R, dengan 5 layer tersembunyi (hidden layer) dengan Training function 500 dan laju pembelajaran (α) 0.4. Hasil klasifikasi untuk data uji menunjukkan bahwa kondisi jantung normal mendapatkan hasil akurasi sebesar 97.73%, supraventricular arrhythmia sebanyak 79.41%, ventricular tachyarrhythmia sebesar 100% dan atrial fibrillation sebanyak 92%, sehingga akurasi rata-rata didapatkan sebesar 93.99%. Setelah itu dilakukan cross validation dan didapatkan akurasi rata-rata adalah 93.08%.
=======================================================================================================
The physiological conditions of normal and abnormal human heart (arrhythmia) can be known by EKG signals (Electrocardiography). This study aims to perform the introduction and classification of the human heart condition. These heart conditions are normal sinus rhythm, supraventricular arrhythmia, ventricular tachyarrhythmia and atrial fibrillation. Modeling the signal using Pan-Tompkin algorithm, where by using the algorithm, we can extract the important information on signal. The obtained parameters are used as input for Backpropagation (multi layer perceptron) with function as training data and test data. The data used by each subject is 10 seconds with a total of 69 subjects representing 4 output classes. The test results show that the best classification uses 4 input parameters ie complex QRS, RR interval, bpm and high Q-R values, with 5 hidden layers with Training function 500 and learning rate (α) 0.4. The results of the classification for the testing data showed for normal heart condition 97.73%, supraventricular arrhythmia 79.41%, 100% ventricular tachyarrhythmia and atrial fibrillation of 92%, so the average accuracy was 93.99%. After that cross validation and got the average accuracy is 93.08%.
| Item Type: | Thesis (Masters) |
|---|---|
| Additional Information: | RTE 616.120 754 7 Oct k-1 3100018074121 |
| Uncontrolled Keywords: | EKG; Arrhythmia; klasifikasi sinyal jantung; Backpropagation; Pan-Tompkin Algorithm |
| Subjects: | R Medicine > RC Internal medicine > RC683.5.E5 Electrocardiography T Technology > TK Electrical engineering. Electronics Nuclear engineering T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105.546 Computer algorithms |
| Divisions: | Faculty of Electrical Technology > Electrical Engineering > 20101-(S2) Master Thesis |
| Depositing User: | Vika Octaviani |
| Date Deposited: | 06 Feb 2018 02:27 |
| Last Modified: | 30 Apr 2020 02:25 |
| URI: | http://repository.its.ac.id/id/eprint/49236 |
Actions (login required)
 |
View Item |