| Yazarlar (2) |
Fatih Murathan Yılmaz
|
Doç. Dr. Mustafa Cem KASAPBAŞI
İstanbul Ticaret Üniversitesi |
| Özet |
| About %1 of the whole population of the world which constitutes more than 50 million people are affected by epilepsy and epileptic seizures (Litt, Echauz 2002) (Kandel ve ark., 2000). Epileptic seizures are caused by a disturbance in the electrical activity of the brain. Detecting epileptic seizure is generally carried out by the expert opinion after examining the electroencephalographic (EEG) signal. This is a manual process and heavily relies on the expertise of the physician. Therefore automated diagnosis or aiding systems are required to assist physicians to diagnose with fewer errors. In this study, a well known (Andrzejak et al. 2001) dataset is used for classifying the existence of epileptic seizures. Different configurations of the data set have been studied with many data mining and machine learning algorithms in the literatüre, some of which are Logistic Regression, Wavelet Method, Decision Tree, Support Vector Machine, Dense Neural networks, etc.. In this study, a classification model was developed by using Random Forest to meet the good diagnosis expectation, and results were compared with different methods studied on the same data set. In some cases of the studied experiments above 99,78 percent of accuracy, 99,95% specificity, and 99,61% sensitivity are obtained, indicating a good sign of classification model. |
| Anahtar Kelimeler |
| Makale Türü | Özgün Makale |
| Makale Alt Türü | Diğer hakemli ulusal dergilerde yayınlanan tam makale |
| Dergi Adı | İstanbul Ticaret Üniversitesi Teknoloji ve Uygulamalı Bilimler Dergisi |
| Dergi ISSN | 2645-8969 |
| Makale Dili | Türkçe |
| Basım Tarihi | 02-2021 |
| Cilt No | 3 |
| Sayı | 2 |
| Makale Linki | https://dergipark.org.tr/tr/pub/icujtas/issue/57754/884801 |
