Dual-modal Physiological Feature Fusion-based Sleep Recognition Using CFS and RF Algorithm

Research has demonstrated a significant overlap between sleep issues and other medical conditions. In this paper, we consider mild difficulty in falling asleep (MDFA). Recognition of MDFA has the potential to assist in the provision of appropriate treatment plans for both sleep issues and related medical conditions. An issue in the diagnosis of MDFA lies in subjectivity. To address this issue, a decision support tool based on dual-modal physiological feature fusion which is able to automatically identify MDFA is proposed in this study. Special attention is given to the problem of how to extract candidate features and fuse dual-modal features. Following the identification of the optimal feature set, this study considers the correlations between each feature and class and evaluates correlations between the inter-modality features. Finally, the recognition accuracy was measured using 10-fold cross validation. The experimental results for our method demonstrate improved performance. The highest recognition rate of MDFA using the optimal feature set can reach 96.22%. Based on the results of current study, the authors will, in projected future research, develop a real-time MDFA recognition system.