Association Rule-Based Analysis of Contributing Factors to Sleep Disorders

Abstract

This study employs the Apriori algorithm in association rule analysis to investigate influencing factors of sleep disorders, with the ultimate goal of providing actionable sleep recommendations. The methodology consists of three key phases: first, dataset preprocessing involving discretization of continuous variables. Second, chi-square testing was used to select five significant variables as independent factors. Finally, association rule mining using the Apriori algorithm to extract meaningful patterns. The findings reveal three key patterns: first, younger populations show significantly elevated insomnia risk with 0.471 support and 0.83 confidence, and this risk persists even with sufficient sleep duration, as evidenced by 0.906 confidence. Second, middle-aged women demonstrate greater susceptibility to sleep apnea, supported by a combined rule confidence of 0.938. Third, younger males exhibit better sleep health outcomes with 0.607 support, while low stress levels show a strong positive association with normal sleep patterns, yielding a 1.205 lift. Additionally, the observed association between moderately high physical activity levels and insomnia, though showing 0.946 confidence, may be influenced by confounding factors. Consequently, insomnia management in younger populations should focus on stress reduction and lifestyle modifications, while middle-aged women with sleep apnea may benefit from therapies such as non-invasive positive pressure ventilation. For the general population, public education initiatives should be implemented to enhance overall sleep health literacy.