Predicting Obesity Levels with High Accuracy: Insights from a CatBoost Machine Learning Model
Authors:
Open Access
- 9 May 2024
- journal article
- research article
- Published by PT. Heca Sentra Analitika in Infolitika Journal of Data Science
- Vol. 2 (1) , 17-27
Abstract
This study aims to develop a machine learning model using the CatBoost algorithm to predict obesity based on demographic, lifestyle, and health-related features and compare its performance with other machine learning algorithms. The dataset used in this study, containing information on 2,111 individuals from Mexico, Peru, and Colombia, was used to train and evaluate the CatBoost model. The dataset included gender, age, height, weight, eating habits, physical activity levels, and family history of obesity. The model's performance was assessed using accuracy, precision, recall, and F1-score and compared to logistic regression, K-nearest neighbors (KNN), random forest, and naive Bayes algorithms. Feature importance analysis was conducted to identify the most influential factors in predicting obesity levels. The results indicate that the CatBoost model achieved the highest accuracy at 95.98%, surpassing other models. Furthermore, the CatBoost model demonstrated superior precision (96.08%), recall (95.98%), and F1-score (96.00%). The confusion matrix revealed that the model accurately predicted the majority of instances in each obesity level category. Feature importance analysis identified weight, height, and gender as the most influential factors in predicting obesity levels, followed by dietary habits, physical activity, and family history of overweight. The model's high accuracy, precision, recall, and F1-score and ability to handle categorical variables effectively make it a valuable tool for obesity risk assessment and classification. The insights gained from the feature importance analysis can guide the development of targeted obesity prevention and management strategies, focusing on modifiable risk factors such as diet and physical activity. While further validation on diverse populations is necessary, the CatBoost model's results demonstrate its potential to support clinical decision-making and inform public health initiatives in the fight against the global obesity epidemic.Keywords
This publication has 33 references indexed in Scilit:
- A review of the application of deep learning in obesity: From early prediction aid to advanced management assistanceDiabetes & Metabolic Syndrome: Clinical Research & Reviews, 2024
- Obesity-Associated Comorbidities: Health Consequences2022
- Effect of interaction between obesity-promoting genetic variants and behavioral factors on the risk of obese phenotypesMolecular Genetics and Genomics, 2021
- Dataset for estimation of obesity levels based on eating habits and physical condition in individuals from Colombia, Peru and MexicoData in Brief, 2019
- Socio-demographic inequalities in cigarette smoking in Indonesia, 2007 to 2014Preventive Medicine, 2019
- Energy balance and obesity: what are the main drivers?Cancer Causes & Control, 2017
- Field evaluation of a random forest activity classifier for wrist-worn accelerometer dataJournal of Science and Medicine in Sport, 2017
- Global, regional, and national prevalence of overweight and obesity in children and adults during 1980–2013: a systematic analysis for the Global Burden of Disease Study 2013The Lancet, 2014
- Sedentary Behavior and Dietary Intake in Children, Adolescents, and AdultsAmerican Journal of Preventive Medicine, 2011
- Sedentary Behaviors and Subsequent Health Outcomes in AdultsAmerican Journal of Preventive Medicine, 2011