To create a risk prediction model for Obstructive Sleep Apnea (OSA) using an Artificial Immune System (AIS) and Logistic Regression (LR) to enhance diagnostic accessibility through machine learning techniques.
Approach:
Data Collection: Data attributes included PSG attributes, sleep questionnaire data, and physiological characteristics to assess OSA severity using the Apnea-Hypopnea Index (AHI), which quantifies the severity of OSA.
Key Findings:
OSA is under-diagnosed due to the limitations of traditional diagnostic tools like polysomnography (PSG).
Home Sleep Testing (HST) offers a less intrusive and cost-effective alternative but has limited physiological indicators.
Combining HST with questionnaires and physiological parameters can enhance diagnostic accuracy.
Interpretation:
The study highlights the need for accessible diagnostic tools for OSA, leveraging machine learning techniques to improve early detection.
Limitations:
The study's findings are based on a specific population in Taiwan, which may limit generalizability due to cultural and environmental factors.
Reliance on self-reported symptoms and questionnaire data may introduce bias in the assessment of OSA.
Conclusion:
The proposed model aims to facilitate early diagnosis of OSA using easily obtainable clinical data.
