SVM-based IADL score correlation and classification with EEG/ECG signals

This paper explores the correlation between the subjective IADL assessment and the objective EEG/ECG signals measurement. Thirty elderly participants are scored by IADL and classified into three groups, that is, the high score, the medium score and the low score groups, and each participant's collected EEG/ECG signals is then attributed to the groups correspondingly. Six equations of extraction methods, including five for EEG and one for ECG, are applied to the EEG/ECG signals from each participant. Thereafter, the extracted features are trained by SVM and classified by one-against-all method in terms of group. The experiment is shown that 82% of accuracy can be reached by the proposed extracted methods and the proposed framework.