Analysis of the Impact of Lifestyle Habits on the Spread of COVID-19 Using Artificial Intelligence

Pandemics are events that significantly impact the social and economic structures of societies, and identifying the factors influencing their spread provides valuable insights for managing these crises. This study analyzed the behavioral characteristics and habits of individuals infected and not infected during the COVID-19 pandemic in Türkiye using artificial intelligence techniques, specifically classification and association rule methodologies. The findings revealed that the Random Committee algorithm performed effectively. Additionally, feature reduction was applied, and the RRF algorithm achieved higher performance with 81% accuracy and a kappa value of 0.3663. In the subsequent analysis of association rules, 21 rules were identified for the "Yes" class (infected with COVID-19), while 2805 rules were found for the "No" class. The results indicated that individuals who live alone, have 1-3 close contacts per day, and spend 4-6 hours in close contact outside the home exhibited strong associations in the "Yes" class. For the "No" class, individuals who frequently avoid travel, have had no significant weight change since the start of the pandemic, work from home, and sometimes avoid small social gatherings in open spaces showed strong associations. In conclusion, the study scientifically demonstrated that lifestyle habits impact the transmission and spread of pandemics and that these factors can be modeled using artificial intelligence techniques.