Machine Learning Approaches for Obesity Classification and Prediction - An Analysis of Demographic, Lifestyle, and Health Factors

Aug 4, 2026

Upload by Skyler

The study addresses obesity as a major global public health threat linked to chronic diseases and reduced life quality. It investigates demographic, lifestyle, and health determinants of obesity using multiple machine learning classifiers. A curated dataset supports model training and evaluation with Decision Tree, Random Forest, Gradient Boosting, Support Vector Machine, and K-Nearest Neighbors. Results show Random Forest and Gradient Boosting achieve the highest accuracy (95.0% and 95.3%), supporting the role of advanced predictive analytics for targeted intervention.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages5

Reading Time13 min

AccessFree to download

Introduction

Literature Review

Machine Learning Approaches for Obesity Classification and Prediction: An Analysis of Demographic, Lifestyle, and Health Factors Joel Azu Department of Science and Engineering Solent University Southampton, United Kingdom [ijoelazu@gmail.com](ijoelazu@gmail.com) Shakeel Ahmad Department of Science and Engineering Solent University Southampton, United Kingdom [shakeel.ahmad@solent.ac.uk](shakeel.ahmad@solent.ac.uk) Abstract—The high level of obesity poses a serious public health problem across the globe, being a leading cause of various chronic diseases and substantial threats to the quality of people’s lives. Given the ever-growing rates of obesity and its impact on individual well-being, the current study aims to explore numerous factors that drive obesity, using machine learning algorithms to solve classification tasks. By working on a rich dataset that contains a range of demographic, lifestyle, and health parameters, several classifiers were developed and tested, namely Decision Tree, Random Forest, Gradient Boosting, Support Vector Machine, and K-Nearest Neighbors. The resulting outcomes indicated that both Random Forest and Gradient Boosting algorithms were highly accurate in the classification of obesity, with 95.0% and 95.3% accuracy rates, respectively. The obtained results confirm the critical role of various machine learning approaches in understanding obesity and developing predictions for more focused intervention. This study offers considerable input into obesity epidemiology literature and demonstrates the utility of advanced analytical appraisal in public health. Keywords—Obesity Classification, Machine Learning, Ensemble Methods, Public Health, Predictive Analytics I. INTRODUCTION Obesity is an epidemic public health problem that affects people of all ages, races and socioeconomic status. As defined by the World Health Organization [1], obesity is a disorder in which an abnormal or excessive fat accumulation may adversely affect a person’s health. Obesity has many health consequences such as heart diseases, diabetes, cancer etc. [2] . Recent statistics from Public Health England [3] show that 63% of adults in the UK are classified as overweight (half moderate, half severe), which is a statistic, for reasons to be explored momentarily likely mirrored throughout many countries with similar 'developed world'socioeconomic/environmental systems. Among children it is no less alarming, with one in three 10-11 year olds overweight or obese by the time they leave primary school, and nearly a fifth of these falling into the obesity bracket. These statistics highlight the pressing demand for effective public health interventions to address growing disease rates of obesity. Obesity is associated with numerous adverse health events, making the case for obesity to be targeted in both prevention and intervention efforts. In the realm of public Raza Hasan Department of Science and Engineering Solent University Southampton, United Kingdom [raza.hasan@solent.ac.uk](raza.hasan@solent.ac.uk) Salman Mahmood Department of Computer Science Nazeer Hussain University Karachi, Pakistan [salman.mahmood@nhu.edu.pk](salman.mahmood@nhu.edu.pk) discourse, it is also important to carry a nuanced blend between celebrating body positivity and recognizing that medical research has long recognized obesity as an incredibly serious health hazard. Educating people with information about the lifestyle causes of obesity is necessary to enable more informed choices and a healthier lifestyle. In this study, we aimed at exploring the enigma of what determines obesity levels by making use of a rich dataset that covered different demographic, lifestyle and health conditions to serve in clearly evaluating its determinants beyond political preferences. This research seeks to elucidate the interplay of these factors and obesity levels using a comprehensive approach. Finally, this research aims to generate robust machine

FAQs

What problem does the study target, and why is it important?

The study targets obesity as an epidemic public health problem because it contributes to many chronic diseases and threatens people’s quality of life, requiring effective prevention and intervention.

Which machine learning models were used for obesity classification and prediction?

The study developed and tested Decision Tree, Random Forest, Gradient Boosting, Support Vector Machine, and K-Nearest Neighbors models.

Which algorithms performed best, and what accuracy was reported?

Random Forest and Gradient Boosting performed best, with accuracy rates of 95.0% and 95.3% respectively for classifying obesity.

Aug 4, 2026

Upload by Skyler