Sleep Disorder Data Stream Classification Based on Classifiers Ensemble and Active Learning

摘要: Polysomnography (PSG) screening for obstructive sleep apnea (OSA) is time consuming. The OSA classification very important medical scientists and machine learning researchers. In the current work, we developed a method electrocardiogram (ECG) data. data set has two labels: disorders or not. As result, Active Learning used as technique. Data stream in non-stationary environment attaining more attention recently. It highly challenging task, since concept drift limited labeled Therefore, model needed to be struggling with detection need of To solve these issues, propose an efficient semi-supervised this paper which uses detect unsupervised way Classifiers Ensemble keep higher predictions combined weighted majority voting. Experiments results on real-world synthetic data-sets show effectiveness proposed approach. For initial experiment, use existing set. includes every 10 seconds, up 6000 35 patients. We have 80% training purposes 20% testing purposes. that our can effectively OSA. accuracy predicted result 71%. Future research area will obtain from hospitals algorithms prediction.