Abstract
This paper explores the escalating global concern of obesity, emphasizing the significance of identifying high-risk individuals to deploy targeted intervention strategies. Employing the University of California, Irvine (UCI) Machine Learning Repository dataset of 2,111 subjects from diverse regions, the classification of obesity levels was based on the Mexican Normativity, which closely aligns with Centers for Disease Control and Prevention (CDC) standards. The primary objective was to assess the predictive capabilities of an array of machine learning models in forecasting obesity levels based on lifestyle habits, excluding direct parameters like height and weight. An enhanced Logistic regression model, LogitBoost model, Random Forests, XGBoost, Support Vector Machines (SVM), Naive Bayes classifiers, and K-Nearest Neighbors (KNN) were employed for analysis. Through cross-validation, this research determined the hierarchy of factors contributing to obesity, spotlighting variables like ‘Consumption of food between meals’ and ‘Obesity among family members’ as major contributors. The results indicate that while LogitBoost performed optimally among Boost algorithms, its performance was slightly below traditional methods. This study’s unique approach of emphasizing lifestyle predictors, excluding direct height and weight variables, underscores the need for targeted, personalized intervention strategies in managing the global obesity epidemic.
Reference19 articles.
1. Obesity as a medical problem
2. World Health Organization, World Health Organization. https://www.who.int/news-room/fact-sheets/detail/obesity-and-overweight, (2021).
3. What use is the BMI?
4. National Health Service, 2023, https://www.nhs.uk/conditions/obesity/.
5. Centers for Disease Control and Prevention, Centers for Disease Control and Prevention. 2022. https://www.cdc.gov/obesity/index.html.