Predictive Modeling for Identifying Insomnia Risk Factors: An Investigative Approach

Show simple item record

dc.contributor.author Herath, H.M.S.S.
dc.contributor.author Herath, H.M.K.K.M.B.
dc.date.accessioned 2024-03-14T10:28:28Z
dc.date.available 2024-03-14T10:28:28Z
dc.date.issued 2023-11-30
dc.identifier.citation Herath, H.M.S.S., & Herath, H.M.K.K.M.B. (2023). Predictive Modeling for Identifying Insomnia Risk Factors: An Investigative Approach. Presented at the 2023 International Research Conference of Sri Lanka Technology Campus, Colombo, Sri Lanka, December 14th-15th. [ORCID IDs: 0009-0008-9702-5576 (H.M.S.S. Herath), 0000-0002-1873-768X (H.M.K.K.M.B. Herath)] en_US
dc.identifier.isbn 978-624-6045-02-9
dc.identifier.uri http://repo.sltc.ac.lk/${dspace.ui}/handle/1/403
dc.description.abstract Global populations are significantly impacted by insomnia, a prevalent sleep problem that negatively impacts daily functioning and general well-being. The intricacies of insomnia are explored in this study by utilizing a large dataset that includes both self-reported tests and thorough questionnaires covering various topics, including sleep habits, stress levels, early life events, and cognitive impairments. The study's main objectives are finding relevant components, examining correlations, and utilizing predictive modeling approaches to reveal important insights. We used advanced feature selection techniques to understand the complex interactions between variables. This study examined the intricacies of insomnia's effects on adolescents utilizing a range of statistical metrics, including correlation coefficients and pvalues. P-values, which show how significant the observed links are, and correlation coefficients, which show how strong and which way the relationships are going, are important metrics in our analysis. Using a variety of machine learning methods, such as Decision Trees (DT), k-Nearest Neighbors (k-NN), Support Vector Machines (SVM), and Naive Bayes (NB), one of the study's main goals was to predict insomnia-related outcomes. Among the models evaluated, the Decision Tree classifier was the most accurate, with an exceptional accuracy rate of 89.47% for both feature selection strategies. These results highlight how reliable Decision Trees are at identifying patterns of sleeplessness. Additionally, the investigation found statistically significant correlations between particular demographic characteristics and insomnia. An important positive link between sex and insomnia was found, with a correlation coefficient of 0.078 and a p-value of 0.001. Age and insomnia showed a significant positive link (correlation coefficient = 0.250). However, the p-value of 0.553 suggests that more research is needed to understand this relationship fully. Further supporting the need to consider these factors for a thorough understanding and management of insomnia, the study found significant correlations between race (correlation coefficient = 0.05, p-value = 0.0) and ethnicity (correlation coefficient = 0.179, p-value = 0.716) with insomnia en_US
dc.language.iso en en_US
dc.publisher Sri Lanka Technological Campus en_US
dc.subject Insomnia en_US
dc.subject Machine learning en_US
dc.subject Predictive modeling en_US
dc.subject Sleep disorders en_US
dc.subject Sleep patterns en_US
dc.title Predictive Modeling for Identifying Insomnia Risk Factors: An Investigative Approach en_US
dc.type Article en_US


Files in this item

This item appears in the following Collection(s)

Show simple item record

Search SLTC e-Repository


Browse

My Account