Eating healthier: Exploring nutrition information for healthier recipe recommendation

•We propose a healthy recipe recommendation framework (NutRec), which first builds a healthy pseudo-recipe considering the nutritional values and then scans the recipe dataset for items resembling the pseudo-recipe. Our proposed NutRec relies not only on the relations between the ingredients themselves but also on those of their quantities, which ultimately dictate the healthiness of a recipe. To the best of our knowledge, no prior study has incorporated these features.•The pseudo-recipe is a list of ingredients with their quantities, and the nutritional values of the pseudo-recipe should match the predefined targets as best as possible. To generate the pseudo-recipe, we first propose an embedding-based ingredient predictor, which embeds all the ingredients into a latent space and predicts the supplemented ingredients based on the distances of ingredient representations; we then propose an amount predictor to compute the quantities of the supplemented ingredients.•We conduct extensive experiments with two real recipe datasets, and the experimental results confirm the superiority of our methods over the baselines. To facilitate the community research, we have publicly released the datasets. With the booming of personalized recipe sharing networks (e.g., Yummly), a deluge of recipes from different cuisines could be obtained easily. In this paper, we aim to solve a problem which many home-cooks encounter when searching for recipes online. Namely, finding recipes which best fit a handy set of ingredients while at the same time follow healthy eating guidelines. This task is especially difficult since the lions share of online recipes have been shown to be unhealthy. In this paper we propose a novel framework named NutRec, which models the interactions between ingredients and their proportions within recipes for the purpose of offering healthy recommendation. Specifically, NutRec consists of three main components: 1) using an embedding-based ingredient predictor to predict the relevant ingredients with user-defined initial ingredients, 2) predicting the amounts of the relevant ingredients with a multi-layer perceptron-based network, 3) creating a healthy pseudo-recipe with a list of ingredients and their amounts according to the nutritional information and recommending the top similar recipes with the pseudo-recipe. We conduct the experiments on two recipe datasets, including Allrecipes with 36,429 recipes and Yummly with 89,413 recipes, respectively. The empiri