Deriving a recipe similarity measure for recommending healthful meals

A recipe recommender system may stimulate healthful and varied eating, when the presented recipes fit the lifestyle of the user. As consumers face the barrier to change their eating and cooking behavior, we aim for a strategy to provide more healthful variations to routine recipes. In this paper, a similarity measure for recipes is derived by taking a user-centered approach. Such a measure can be used to recommend healthier alternatives to commonly selected meals, which are perceived to be similar. Recipes presented using this strategy may fit the demand for health and variation within the boundaries of a busy lifestyle. Having derived and evaluated a recipe similarity measure, we explore its use through an at-home trial.