Scholarlink

<p>The world has undergone significant transformations in the past decade, largely influenced by technological advancements and the emergence of new, widespread pandemics. As a result, human lives have experienced numerous changes with both positive and negative impacts on health. Increased virtual engagements, the prevalence of diseases, reduced physical activity, excessive use of medications, and changes in biological behavior contribute to various health-related issues, including insomnia—a recognized sleep disorder. Instances of this disorder can manifest independently or arise as a consequence of sleep-related problems induced by the factors above. Expensive tests and equipment are also not available in many developing countries. To address this disparity, our solution involves developing an intelligent model utilizing a machine-learning approach to predict chronic insomnia. Employing six distinct machine learning classifiers, it is observed that the Random Forest classifier consistently outperformed the others across various socio-demographic, biological, and oxygen saturation (SpO<sub>2</sub>) support labels. Accordingly, the generated model diagnoses cases based on features in the dataset, using a modified K-Nearest Neighbors approach enhanced by an n-Random Forest model. This integration achieves high accuracy in identifying cases during pandemics. Notably, the Employment-Status Label model achieved a remarkable accuracy of 99%, enabling the accurate classification of individuals with insomnia. Additionally, the Oxygen Saturation Support Label model exhibited an accuracy of 98% in identifying the insomniac class.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Analyzing the impact of the pandemic on insomnia prediction using machine learning classifiers across demographic groups

  • 期刊论文

  • 发表日期：

    2025年6月24日

查看全文
  • Soran R. Mohammed-Taha,
  • Aso M. Aladdin,
  • Araz Ibrahim Mustafa,
  • Dler O. Hasan,
  • Rebwar Khalid Muhammed,
  • Tarik A. Rashid

摘要

The world has undergone significant transformations in the past decade, largely influenced by technological advancements and the emergence of new, widespread pandemics. As a result, human lives have experienced numerous changes with both positive and negative impacts on health. Increased virtual engagements, the prevalence of diseases, reduced physical activity, excessive use of medications, and changes in biological behavior contribute to various health-related issues, including insomnia—a recognized sleep disorder. Instances of this disorder can manifest independently or arise as a consequence of sleep-related problems induced by the factors above. Expensive tests and equipment are also not available in many developing countries. To address this disparity, our solution involves developing an intelligent model utilizing a machine-learning approach to predict chronic insomnia. Employing six distinct machine learning classifiers, it is observed that the Random Forest classifier consistently outperformed the others across various socio-demographic, biological, and oxygen saturation (SpO2) support labels. Accordingly, the generated model diagnoses cases based on features in the dataset, using a modified K-Nearest Neighbors approach enhanced by an n-Random Forest model. This integration achieves high accuracy in identifying cases during pandemics. Notably, the Employment-Status Label model achieved a remarkable accuracy of 99%, enabling the accurate classification of individuals with insomnia. Additionally, the Oxygen Saturation Support Label model exhibited an accuracy of 98% in identifying the insomniac class.