Machine Learning-Based Caloric Expenditure Prediction for Personalized Fitness Assessment during Physical Activity

The goal of this project is to use machine learning for parsing how our body works when burning calories in different activities. Data was taken consisting of the variables, age, gender, height, weight to record body metrics also how long participants had engaged in exercise. The heart rate and temperature were taken as well. Once this data is preprocessed and arranged, we use this data to train multiple machine learning models such as Linear Regression, Decision Trees etc. The metrics used to evaluate these models — Mean Squared Error, R2 Score etc. indicate that the Random Forest Regressor is the best model in predicting caloric expenditure with minimum error of accuracy. To achieve this, we create a Streamlined UI application (below) that allows the users to input their data age, gender, height and weight in metric units of [kg], type of exercise hour duration — walking or running— heart rate during training and temperature in Celsius unit. The application then uses the learned machine learning model to predict how many calories were burned and produce the output with the accuracy level of 93.77%.