Obstructive sleep apnea-hypopnea syndrome (OSAHS) is an increasingly significant health concern due to its widespread occurrence, heightened risk factors, and significant mortality rates. To address this issue, researchers have focused on utilizing the blood oxygen saturation (SpO2) signal to analyze the occurrences of apnea or hypoventilation episodes during sleep, resulting in the development of the apnea-hypopnea index (AHI). In order to extract relevant information from the SpO2 signals, 35 Time Domain characteristics have been identified and studied. To enhance the practical application of these features in Industry 5.0 supply chains, a feature selection procedure was implemented, effectively reducing the dimensions from 7 to just 5. This reduction has significantly improved the feasibility of integrating these features into real-world scenarios. To select the most relevant features, a combination of the competitive swarm optimizer algorithm and the Pearson correlation coefficient was utilized in this study. By employing these techniques, the researchers were able to identify the most informative features for analysis and classification. The study achieved promising results using a random forest classifier, reporting an accuracy rate of 86.92% and a specificity rate of 90.7%. These findings highlight the potential of using the selected features for the accurate detection and diagnosis of OSAHS, contributing to the development of effective treatment strategies and improved patient outcomes.
