Early Obesity Risk Assessment Using Machine Learning to Analyse Lifestyle and Demographic Data

Obesity has emerged as a major worldwide health issue, linked to several illnesses including diabetes, heart problems, and several types of cancer. Predicting obesity levels in advance may aid in tailored healthcare plans and prompt action. Based on a variety of demographic, behavioral, and lifestyle characteristics, this research uses machine learning algorithms Random Forest, Gaussian Naive Bayes, and Decision Tree to predict obesity levels. The dataset includes variables that affect body weight, such as eating habits, physical activity levels, and other factors, in addition to characteristics like "Gender," "Age," "Height," "Weight," and family history of obesity. Target labels classify people into seven different weight classes: 'Normal Weight', 'Overweight Level I', 'Overweight Level II', 'Obesity Type I', 'Insufficient Weight', 'Obesity Type II', and 'Obesity Type III'. The objective is to properly categorize people depending on their degrees of obesity by training models with these input characteristics. The purpose of the research is to evaluate how well various algorithms predict these categories and whether or not these predictions match actual trends in obesity. Metrics like accuracy, precision, recall, and F1-score are used to assess each model's performance. The best effective algorithm for predicting obesity is found via comparative analysis, with an emphasis on optimizing predictive accuracy. The findings of this study may be used to incorporate machine learning into healthcare apps to provide early risk assessments for obesity, assisting in the avoidance of associated health problems.