Predictive personalized workout and dietary guidance system

Chua, Alicia Xiu Wen (2025) Predictive personalized workout and dietary guidance system. Final Year Project, UTAR.

PDF Download (7Mb)

Abstract

The COVID-19 pandemic has increased the use of fitness and dietary mobile applications. While existing fitness and dietary applications offer useful functionalities, they often fail to deliver personalised recommendation that account for individual differences. This project proposes the development of the “Predictive Personalised Workout and Dietary Guidance System”, a comprehensive mobile application designed to address the shortcomings of existing systems. This application utilises publicly available datasets and integrates artificial intelligence to analyse user data such as weight, height, gender, and age, offering tailored recommendations that evolve with user progress. Deep learning models were integrated and evaluated to predict users’ Body Mass Index (BMI) classification. During the development phase in, three predictive models were implemented and evaluated: a Deep Neural Network (DNN), a U-Net-based Convolutional Neural Network (CNN) and a Random Forest. Among them, the CNN model achieved the highest test accuracy of 90.36%, but DNN and Random Forest only achieved a test accuracy of 88.70% and 85.00%, respectively, proving the U-Netbased CNN model is more effective and reliable for BMI classification. This result highlights the advantage of using a U-Net-based CNN architecture for personalised health predictions within the application. Unlike existing systems, which often focus primarily on exercise tracking with minimal dietary support and lack of suitable workout recommendations. The application will include functions such as fitness and dietary tracking, community platform to enhance user engagement and motivation, artificial intelligence (AI) chatbot that support users with personalised guidance, and weight tracking function. In addition, the system implements a personalised dietary module that uses AI to analyse meals’ macronutrient intake, enabling users to adopt a more health-oriented diet. With the comprehensive functions offered by the system, users are expected to benefit from more precise and adaptable health guidance, thereby improving long-term commitment and overall health outcomes.

Actions (login required)