Predict Overweight or Obesity using Body Fat Percentage with Machine Learning Classification Algorithms to Enhance Accuracy
N Nalini, Subuddi Nagaraju, R. Jayanthi, Chhotu Ram, G Guna., Deepak Pandey
Abstract
In today's world, being overweight has become a serious global health problem. This obesity represents an excessive or unusual buildup of body fat, often resulting from unhealthy lifestyle choices. Factors such as frequent consumption of junk food, irregular sleep patterns, and prolonged sedentary behavior contribute significantly to this issue. Adolescents, in particular, are highly vulnerable due to a lack of awareness about their habits. Obesity is recognized as a complex medical condition that increases the likelihood of severe health complications, including heart disease, stroke, and liver cancer. One of the major objectives of this study is to warn people about their risk of becoming obese and identify its underlying causes. The collected data from over 1,100 individuals among different kind of age groups, including both obese and non-obese participants. Performance of these classifiers was evaluated using key performance metrics. Obesity levels were classified as high, medium, and low depending on the findings of the experiment. The Gradient Boosting approach fared the worst compared to all the models, with the lowest metric scores and an accuracy of 96%, while accuracy of Logistic Regression 93%.