Download PDFOpen PDF in browserUtilizing AI to Tackle Child Overnutrition in MoroccoEasyChair Preprint 149177 pages•Date: September 17, 2024AbstractThis study investigates the predictive power of various machine learning classifiers in identifying overnutrition among children under five years old in Morocco. Using data from the 2018 national population and family health survey, the research employs ten machine learning algorithms: Logistic Regression, Support Vector Machines (SVM), Gradient Boosting, Random Forest, XGBoost, k-Nearest Neighbors, Decision Trees, Naive Bayes, Artificial Neural Networks, and Deep Learning models. The performance of these models was assessed using accuracy, sensitivity, specificity, kappa statistic, and area under the curve (AUC). The results reveal that Logistic Regression and SVM were the most effective models, achieving nearly ninety percent accuracy for predicting overweight and approximately ninety-seven percent accuracy for predicting obesity. These models also demonstrated high sensitivity and specificity. Gradient Boosting and Random Forest also showed strong performance, while Naive Bayes, despite its lower accuracy, excelled in AUC, indicating its proficiency in overall class distinction. The findings highlight the significant roles of birth weight, socioeconomic status, and parental education in influencing childhood overnutrition. By focusing on the Moroccan context, this study addresses a gap in the existing literature and provides actionable insights for developing targeted public health interventions. The study underscores the effectiveness of Logistic Regression and SVM in handling complex datasets and the importance of early life factors, socioeconomic status, and parental education. The research suggests the need for incorporating more recent and longitudinal data, exploring a broader range of variables, and employing advanced machine learning techniques in future studies to enhance predictive accuracy into the determinants of childhood overnutrition. Keyphrases: Determinants, Machine Learning Algorithms, childhood obesity, childhood overweight
|