Obesity Risk Prediction Using Machine Learning Approach

Approximately about two billion peoples are affected by obesity that has drawn significant attention on social media. As the sedentary lifestyle which includes consumption of junk foods, no physical activities,spending more on screen,etc are one of the causes of obesity.Obesity generally refers to that a person's body possessing an excessive amount of fat.There is a huge increase in obesity cases which resulting cardiac problems,stroke,insomnia, breathing problems,etc.Type-2 diabetes has been detected in the patients suffering from obesity recently. The studies showing that there are lot of young individuals and children's who has been suffering from overweight and obesity issues in Bangladesh. Here, a strategy for predicting the risk of obesity is proposed that makes use of various machine learning methods. The dataset Obesity and Lifestyle taken from Kaggle site which is collection of different data based on the eating habits and physical conditions,such as height, weight,calorie intake,physical activities are just a few of the 17 different categories in the dataset that reflect the elements that cause obesity. Several machine learning methods include Gradient Boosting Classifier, Adaptive Boosting (ADA boosting), K-nearest Neighbor (K-NN), Support Vector Machine (SVM), Random Forest, and Decision Tree. A few important performance factors are used to group the models. Predicting the levels of high, medium, and low obesity in this case using the experimental results. The gradient boosting techniques have the highest accuracy 97.08% in comparison to other classifiers