Classification and prediction of obesity levels among subjects in Colombia, Peru, and Mexico using unsupervised and supervised learning

• Main Authors: Suhaila, Bahrom, Anuar, Ab Rani, Aisyah Amalina, Mohd Noor
• Format: Article
• Language: English
• Published: Zenodo 2024
• Subjects: QA Mathematics; QA76 Computer software
• Online Access: http://umpir.ump.edu.my/id/eprint/42159/1/Classification%20and%20Prediction%20of%20Obesity%20Levels%20among%20Subjects%20in%20Colombia%2C%20Peru%2C%20and%20Mexico%20Using%20Unsupervised%20and%20Supervised%20Learning.pdf; http://umpir.ump.edu.my/id/eprint/42159/; https://doi.org/10.5281/zenodo.12791087

This research investigates the multifaceted relationship between various factors and obesity rates in Mexico, Peru, and Colombia using a publicly available dataset. Through Python, the study employs classification and clustering analyses, focusing on logistic regression to predict obesity levels and generate actionable recommendations. Combining exploratory data analysis (EDA) and advanced machine learning techniques, the research aims to unveil nuanced insights into obesity determinants. Unsupervised learning methods segmentize individuals, providing deeper insights into obesity profiles. Supervised learning algorithms like logistic regression, random forest, and adaboost classifier predict obesity levels based on labelled datasets, with random forest exhibiting superior performance. The study enhances understanding of obesity classification through machine learning and integrates data inspection, formatting, and exploration using Excel, Python, and graphical user interfaces (GUIs) such as SweetViz and PandaGui. Overall, it offers a comprehensive approach to understanding and addressing obesity using sophisticated analytical tools and methodologies.