Analysis on Various Machine Learning Framework for Obesity Level Prediction

In the realm of public health, predicting obesity levels through machine learning emerges as a crucial initiative. This study employs a dataset comprising 2111 instances and 17 features, encompassing diverse factors such as physical activity, eating habits, and lifestyle attributes. Consideration of various attributes or features is essential in obesity level prediction to capture the intricate and multifaceted nature of the condition. Numerous factors, such as food practices, physical activity, lifestyle decisions, and genetic predispositions, all have an impact on obesity. By incorporating a diverse set of attributes, machine learning models can discern complex patterns and relationships within the data, enabling a more comprehensive and accurate prediction of obesity levels. Utilizing advanced machine learning algorithms, the predictive model achieved a notable accuracy of 97.2% with Multilayer Perceptron ANN Model, followed closely by Gradient Boosting classifiers at 96.2%. Such predictive accuracy is instrumental in developing targeted preventive strategies and timely interventions. Flask based web application has been implemented for real time prediction using MLP model. The obesity scenario in India, compounded by factors like suboptimal dietary habits and sedentary lifestyles, underscores the urgency of effective prediction models for early detection and tailored health interventions.