Recipe Recommendation and Management System

Abstract

This document refers to a project that creates a web-based Recipe Recommendation System to help busy users by focusing on delivering a simple function where users can search specific recipes based on available ingredients, preparation time, and meal type. Beyond the function of filtering, the project enhances the user experience by indicating the nutritional value of each recipe and allowing users to save their favourite recipes to be accessed later. A leading principle of design was to achieve an optimal user experience, so all content of the application is designed to exist solely in the application itself (no external browsers). The application was designed using React.js for the front end of the application, with Material UI (MUI) used for the design of the application. The application uses Node.js with Express as the backend service. User authentication of the application is being delivered through JSON Web Tokens (JWT) as a means for users to log in and save recipes. A small machine learning model consisting of Scikit-learn is being used to recommend recipes based on ingredients the user selected, utilizing TF-IDF for vectorization and cosine similarity for the algorithm function to match the user inputted data, ingredients to the data sofa recipes. The goal of this project ultimately addresses the need for increased healthy eating habits by making home cooking easier, less time-consuming, and more personal for modern users.