Risalah, Desanti N. (2018) Klasifikasi Obstructive Sleep Apnea (OSA) Berdasarkan Fitur Statistik dari RR Interval Pada Sinyal ECG - Classification Of Obstructive Sleep Apnea (OSA) Based On Statistical Features Of RR Interval On ECG Signals. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
07211440000030-Undergraduated_Theses.pdf - Accepted Version Download (4MB) | Preview |
Abstract
Obstructive sleep apnea (OSA) merupakan bentuk gangguan umum yang menyebabkan penderitanya mengalami henti napas saat tidur yang dapat memicu munculnya penyakit kardiovaskular jika tidak tertangani dengan benar. Secara umum, pengujian apnea biasanya menggunakan polysomnography (PSG), yang merupakan prosedur standar untuk diagnosis semua gangguan tidur. Namun sebagian besar kasus sleep apnea saat ini tidak terdiagnosis karena masalah biaya yang mahal dan kurang tersedia karena PSG dilakukan di laboratorium khusus, dimana para staf yang bertugas harus bekerja semalaman. Penelitian ini mengusulkan sebuah metode untuk mendeteksi OSA berdasarkan sinyal ECG. Sinyal ECG dapat dimanfaatkan untuk membuat sistem deteksi yang lebih sederhana dan penanganan yang lebih cepat daripada PSG. Data berupa sinyal ECG didapatkan dari Physionet Database dengan menggunakan 12 subjek rekaman berbeda. Data direkam selama 2 malam berturut-turut menggunakan perekaman single-channel ECG. Sinyal tersebut kemudian dilakukan ekstraksi fitur-fitur statistiknya berdasarkan jarak antar puncak R atau RR interval sinyal ECG yang diproses dalam durasi periode yang singkat. Hasil ekstraksi fitur tersebut digunakan sebagai input klasifikasi menggunakan metode Support Vector Machine untuk dilatih dan diujikan pada rekaman apnea dan non-apnea dari subjek positif dan negatif OSA. Dilakukan pula pengujian untuk mengetahui kinerjanya dengan mendapatkan nilai Accuracy, Sensitivity, dan Specificity dari sistem. Hasil menunjukkan bahwa sistem yang telah dibuat dapat mendeteksi periode gangguan OSA dengan akurasi yang cukup tinggi, yakni sekitar 83.67%. Kedepannya diharapkan sistem yang telah dibuat dapat terus dikembangkan dan dapat digunakan sebagai dasar pengembangan alat deteksi OSA di kemudian hari.
===========================================================================================================
Obstructive sleep apnea (OSA) is a common disorder that person stop breathing during sleep that can cause cardiovascular disease if it is not handled well. Most of sleep apnea cases were tested with polysomnography (PSG), which is the standard procedure for diagnose all sleep disorders. But most of sleep apnea cases are currently undiagnosed because of expenses and inconvenient, because PSG is done at sleep labs, where an expert human observer is needed to work overnight. This research proposes a method for detecting OSA based on ECG signal. ECG signals can be utilized to make detection system simpler and handling faster than PSG. The ECG signals that used are available from Physionet Database with 12 different records. Data is recorded for 2 consecutive nights using ECG single-channel recording. Statistical features are extracted based on RR intervals which processes short duration epochs of ECGs. Features extraction results are used as classification input using Support Vector Machine method to be trained and tested on sleep apnea records from subjects with and without OSA. System could determine its performance by obtaining Accuracy, Sensitivity, and Specificity score. The results show that the system that has been created can recognize epochs of OSA with a high degree of accuracy, approximately 83.67%. In the future, it is expected that the created system can be developed and can be used as basis for future development of a tool for OSA screening.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Additional Information: | RSKom 621.382 2 Ris k-1 3100018076393 |
| Uncontrolled Keywords: | ECG, OSA, RR interval, Support Vector Machine, |
| Subjects: | T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5102.9 Signal processing. |
| Divisions: | Faculty of Electrical Technology > Computer Engineering > 90243-(S1) Undergraduate Thesis |
| Depositing User: | Desanti Nurma Risalah |
| Date Deposited: | 08 Jan 2019 07:35 |
| Last Modified: | 15 Feb 2021 04:41 |
| URI: | http://repository.its.ac.id/id/eprint/58404 |
Actions (login required)
 |
View Item |