The world wastes approximately 2.5 billion tons of food every year. This project aims to mitigate food waste by suggesting recipes based on the ingredients users have at home. The system prioritizes using items that are close to their best-before dates to reduce food waste and help users save money, while also taking users' preferred cuisines into account.

To achieve this, we compared different algorithms (KNN, Linear SVC, Random Forest) to develop a cuisine predictor for a given set of ingredients. We decided to use Linear SVC after cross-validation. Additionally, we calculated the correlation between different cuisines based on the frequency of ingredients used in their recipes. Using this correlation, we defined a distance function between cuisines and obtained a dendrogram, which allowed us to cluster cuisines methodologically. We integrated these components to develop our software.