Open AccessShallow Classifier with Sampling for Sleep Stage Classification of Autism Patients
Author(s) -
Intan Nurma Yulita,
Mohamad Ivan Fanany,
Aniati Murni Arymurthy
Publication year - 2018
Publication title -
international journal of engineering and technology
Language(s) - English
Resource type - Journals
ISSN - 2227-524X
DOI - 10.14419/ijet.v7i4.44.26982
Subject(s) - naive bayes classifier , artificial intelligence , resampling , computer science , perceptron , classifier (uml) , multilayer perceptron , decision tree , polysomnography , pattern recognition (psychology) , data classification , machine learning , artificial neural network , psychology , support vector machine , electroencephalography , psychiatry
Autism is a brain development disorder that affects the patient's ability to communicate and interact with others. Most people with autism get sleep disorders. But they have some difficulty to communicate, so this problem is getting worse. The alternative that can be done is to detect sleep disorders through polysomnography. One of the test purposes is to classify the sleep stages. The doctors need a long time to process it. This paper presents an automatic sleep stage classification. The classification was based on the shallow classifiers, namely naive Bayes, k-nearest neighbor (KNN), multi-layer perceptron (MLP), and C4.5 (a type of decision tree). On the other hand, this dataset has a class imbalance problem. As a solution, this study carried out the mechanism of resampling. The results show that the use of d as a measure of the uniformity of data distribution greatly influenced the classification performance. The higher d, the more uniform the distribution of data (0 <= d <= 1). The performance with d = 1 was higher than d = 0. On the other hand, KNN was the best classifier. The highest accuracy and F-measure were 83.07 and 82.80 respectively.