putri, Nomala Gema Puji (2018) Klasifikasi Epilepsi Kondisi Kejang dan Tidak Kejang berdasarkan Gelombang Otak EEG Menggunakan Empirical Mode Decomposition (EMD) dengan Ektraksi Fitur Bandwidth. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
Text
07211440000038-Undergraduate_Theses.pdf - Accepted Version Download (3MB) |
Abstract
Penelitian melakukan klasifikasi kondisi kejang penderita epilepsi . Hasil dari pemeriksaan berupa sinyal electroencephalogram (EEG) akan digunakan sebagai data untuk melakukan klasifikasi. Data rekaman EEG yang digunakan berasal dari “Klinik für Epileptologie, Universität Bonn” dengan total data 500 dan dari Rumah Sakit Universitas Airlangga dengan total data 500. Data EEG yang diperoleh kemudian didekomposisi menggunakan metode Empirical Mode Decomposition (EMD) menghasilkan Intrinsic Mode function (IMF). IMF yang didapatkan dianalisis menggunakan transformasi Hilbert. Hasil analisis IMF akan digunakan untuk mendapatkan ekstraksi fitur bandwidth yaitu bandwidth modulasi amplitudo (BAM) dan bandwidth modulasi frekuensi (BFM) . BAM dan BFM yang dihasilkan akan menjadi input pada proses klasifikasi menggunakan Support Vector Machine (SVM). Dari hasil penelitian didapatkan rata-rata akurasi terbaik sebesar 96.8% pada data publik dan 96.5% pada data RSUA.
=====================================================================================================
Some research,about clasification of seizure conditions. Epilepsy is one of the brain neurological disorders, caused by functional or structural damage, to determine electroencephalography is used in the evaluation of brain disorders. The result of evaluation electroencephalogram will be used to classify seizure. An EEG dataset, which is publicly in “Clinical für Epileptologie, Universität Bonn” the dataset with 500 EEG dataset and from Airlangga University Hospital Surabaya dataset with 500 EEG dataset obtained by empirical mode decomposition to decompose signals into Intrinsic Mode Functions (IMF's), then processed by Hilbert Transform to produce corresponding analytic signals for each of the inherent modes. The result of IMF analysis by using Hilbert transform will be used to get feature extraction bandwidth namely amplitude modulation bandwidth (BAM) and frequency modulation bandwidth (BFM). The features of BAM and BFM are used as input of the Support Vector Machine (SVM) classifier for seizure and non seizure EEG. From the research, the best accuracy average from public data is 96.8% and best average accuracy on RSUA data is 96.5%.
Item Type: | Thesis (Undergraduate) |
---|---|
Uncontrolled Keywords: | Epilepsi ,EEG, Empirical Mode Decomposition, Pembelajaran Mesin |
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: | Nomala Gema Puji Putri |
Date Deposited: | 18 Jun 2021 12:48 |
Last Modified: | 18 Jun 2021 12:48 |
URI: | http://repository.its.ac.id/id/eprint/58668 |
Actions (login required)
View Item |