Decoding Sleep: Leveraging Machine Learning for Precision Insomnia Classification
Author(s):Salil Choudhary1, Tushar Kheterpal2, Shivam Verma3, Manoj Kumar4
Affiliation: 1,2,3,4 Department of Computer Engineering, Delhi Technological University, Delhi, India ,?Professor, Department of Computer Engineering, Delhi Technological University, Delhi, India
Page No: 1-14
Volume issue & Publishing Year: Volume 2 Issue 5 ,May-2025
Journal: International Journal of Advanced Multidisciplinary Application.(IJAMA)
ISSN NO: 3048-9350
DOI: https://doi.org/10.5281/zenodo.17519213
Abstract:
Sleep disorders such as insomnia and sleep apnoea are pervasive health conditions that negatively impact both physical well-being and cognitive function. Despite their widespread occurrence, many individuals remain undiagnosed due to the high cost, inconvenience, and limited accessibility of conventional diagnostic techniques like Polysomnography (PSG). This research presents a novel, data-driven approach for the classification of sleep disorders using machine learning algorithms applied to lifestyle and health-related data. The study explores the performance of individual classifiers, including Support Vector Machines (SVM), Decision Trees, K-Nearest Neighbours (KNN), Random Forests, and Artificial Neural Networks (ANN), and further enhances predictive accuracy through ensemble learning techniques such as Stacking and Voting classifiers. These ensemble models integrate the strengths of multiple base learners, offering improved generalization and reliability. The methodology involves comprehensive preprocessing, feature engineering, and model optimization to handle the nuances of real-world data. Experimental results demonstrate that ensemble methods significantly outperform traditional models in classification accuracy, precision, recall, and F1-score. By leveraging commonly available health metrics instead of clinical-grade sensor data, the proposed system offers a scalable and cost-effective solution for early diagnosis, particularly suited for remote or resource-constrained settings. This work underscores the potential of machine learning in developing accessible, non-invasive diagnostic tools that support public health initiatives and individual patient care.
Keywords:
Reference:
- Sleep Disorders, Machine Learning, Ensemble Learning, Stacking Classifier, Voting Classifier, Health Data Analytics, Non-Invasive Diagnosis