Turkish Journal of Science and Technology, Journal Year: 2024, Volume and Issue: 20(1), P. 77 - 90
Published: Dec. 3, 2024
The number of people affected by obesity is rising steadily. Diagnosing crucial due to its harmful impacts on human health and it has become one the world’s most important global concerns. Therefore, develop methods that can enable early prediction risk aid in mitigating increasing prevalence obesity. In literature, some rely solely Body Mass Index (BMI) for classification may result inaccurate outcomes. Additionally, more accurate predictions be performed developing machine learning models incorporate additional factors such as individuals’ lifestyle dietary habits, alongside height weight used BMI calculations. this study, potential three different (naive Bayes, decision tree, Random Forest (RF)) predicting levels were investigated. best performance among compared was obtained with RF (accuracy=0.8892, macro average F1-score=0.8618, Macro Average Precision (MAP)=0.8350, Recall (MAR)=0.9122,). addition, feature selection also determine features are significant estimation level. According experimental results selection, method resulted highest score (accuracy=0.9236, MAP=0.9232, MAR=0.9358, F1-score=0.9269) fewer features. demonstrate same dataset enhanced through detailed hyperparameter tuning. Furthermore, applying improve adverse effects irrelevant or redundant degrade model’s effectiveness.
Language: Английский
