Bagusmulya, Aditya (2015) Deteksi Penyakit Epilepsi Berdasarkan Data EEG Otak Manusia Dengan Menggunakan Independent Component Analysis, Wavelet Transform, Dan Multilayer Perceptron. Undergraduate thesis, Institut Technology Sepuluh Nopember.
Preview |
Text
5111100136-Undergraduate Thesis.pdf - Published Version Download (2MB) | Preview |
Abstract
Epilepsi merupakan salah satu kelainan pada otak manusia yang tidak dapat disembuhkan. Penyakit ini menimbulkan kejang pada tubuh dan sangat mengganggu aktivitas. Pada tingkat yang parah, epilepsi dapat membahayakan nyawa penderitanya. Oleh sebab itu, epilepsi harus dideteksi secara dini agar penderita segera mendapatkan penanganan yang tepat sehingga keadaannya tidak memburuk..
Pada Tugas Akhir ini, deteksi epilepsi dilakukan dengan menggunakan beberapa metode, yaitu Independent Component Analysis, Wavelet Transform, dan Multilayer Perceptron. Hasil deteksi diklasifikasikan ke dalam tiga kelas, yaitu normal, epilesi tidak kejang, dan epilepsi kejang. Data rekaman electroencephalogram (EEG) yang digunakan berasal dari ''Klinik für Epileptologie, Universität Bonn” yang diperoleh secara online.
Hasil pendeteksian terbaik dihasilkan dari model yang menggunakan teknik Single Channel Independent Component Analysis pada Independent Component Analysis sebagai
viii
penghilang derau dan ektraksi fitur Discrete Wavelet Transform Daubechies 6 dengan 4 level. Berdasarkan uji coba, metode tersebut menghasilkan akurasi sebesar 92.09%.
========================================================================================================
Epilepsy is one of disorders in human brain that is cannot be healed. This diesease occurs seizuring which bothers patients’ activities. In the worst condition, it endangers patients’ life. Therefore , the epilepsy must be detected since the early beginning so that patients get a proper treatment immediately for avoiding worse condition.
On this undergraduated thesis, epilepsy detection was build by using three methods; The Independent Component Analysis, Wavelet Transform, and Multilayer Perceptron. The result of detection was classified into 3 classes. They were normal, epilepsy non-seizure, and epilepsy seizure. While the electroencephalogram (EEG) record data used was taken from ''Klinik für Epileptologie, Universität Bonn” website.
The best result of classification was achieved by a model that was build by Single Channel Independent Component Analysis technique in Independent Component Analysis, Wavelet Transform, and Multilayer Perceptron as noise removal and Discrete Wavelet Transform using Daubechies 6 with 4 level as feature extraction. Based on test result, the method above obtained an acurracy of 92.09%.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Additional Information: | RSIf 006.32 Bag d |
| Uncontrolled Keywords: | Epilepsi, Independent Component Analysis, Wavelet Transform, Multilayer Perceptron, Klasifikasi |
| Subjects: | Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) |
| Divisions: | Faculty of Information and Communication Technology > Informatics > 55201-(S1) Undergraduate Thesis |
| Depositing User: | Mr. Tondo Indra Nyata |
| Date Deposited: | 13 May 2019 03:44 |
| Last Modified: | 13 May 2019 03:44 |
| URI: | http://repository.its.ac.id/id/eprint/63011 |
Actions (login required)
 |
View Item |