Content-based recommendation directly analyzes dish attributes and matches them against user-declared preferences, making it friendly to new users, but it is prone to filter bubble effects