SVM and Randfores Code for DoA monitoring

According to Etsevo, during anesthesia induction, we have divided the patient’s depth of sedation into awake, light sedation, and deep sedation and collected 90s of EEG and EOG at each sedation depth for predictive analysis. We included 33 patients’ data into the analysis and obtained 297 pieces of data input into the machine learning method for classification and prediction. In order to study whether sound-evoked EOG has an auxiliary effect on anesthesia monitoring, we combine EEG and EOG features or only input EEG features into the random forest and linear support vector machine for classification tasks.