I am building a recipe recommendation engine using Python which takes certain ingredients from the user and suggests some recipes. It uses TF-IDF to vectorize the ingredients. What's special? It is trained on a dataset containing more than 39,000 recipes!
