Web Application Based on Artificial Intelligence for the Control of Healthy Habits in People with Unbalanced Diet in Lima

This study presents the development of a web application designed to promote healthy habits using artificial intelligence, targeting people aged 18 to 50 years in Lima, Peru, who struggle with overweight or unbalanced diets. The application integrates personalized meal plans and exercise routines generated by machine learning algorithms based on users’ health data. The system architecture includes a frontend built with Flutter and a backend using Spring Boot and Java, communicating with a Flask API that processes data with Random Forest models. Data from national health surveys and Kaggle nutrition and exercise databases were used to train the models. The usability of the application was validated through user satisfaction surveys and predictive model performance metrics. The results indicate that the app effectively helps users manage their eating and physical activity habits, with the meals model achieving an accuracy of 92.38%, recall of 93.47%, F1 score of 91.19%, and AUC-ROC of 91.41%, and the exercise model achieving an accuracy of 78.09%, recall of 76.28%, F1 score of 88.23%, and AUC-ROC of 94.90%, thus contributing to healthier lifestyles.