Optimizing MLP Classifier and ECG Features for Sleep Apnea Detection.

Yıl 2015, Cilt: 11 Sayı: 1, 1 - 18, 20.01.2016

Oğuz Timuş Erkan Kıyak

Öz

The purpose of this study is to optimize multilayer perceptron (MLP) classifier and find optimal ECG features to achieve better classification for automated sleep apnea detection. k-fold crossvalidation technique was employed for classification of apneaic events on the apnea database of the DREAMS project containing 12 whole-night Polysomnography (PSG) recordings previously examined by an expert. To achieve the best possible performance with MLP, the correlation feature selection method was utilized. The performance for apnea event diagnosis after optimization of the features and the classifier resulted almost 10% in accuracy, %7 in sensitivity and %13 in specificity.

Anahtar Kelimeler

-

Optimizing MLP Classifier and ECG Features for Sleep Apnea Detection.

Öz

The purpose of this study is to optimize multilayer perceptron (MLP) classifier and find optimal ECG features to achieve better classification for automated sleep apnea detection. K-fold crossvalidation technique was employed for classification of apneaic events on the apnea database of the DREAMS project containing 12 whole-night Polysomnography (PSG) recordings previously examined by an expert. To achieve the best possible performance with MLP, the correlation feature selection method was utilized. The performance for apnea event diagnosis after optimization of the features and the classifier resulted almost 10% in accuracy, %7 in sensitivity and %13 in specificity.

Anahtar Kelimeler

Elektrokardiyogram, kalp atım hızı değişikliği, uyku apnesi, çok katmanlı algılayıcı, sınıflandırma, ilinti öznitelik seçimi