As many people will refer to a recipe when cooking, there are several recipe-sharing websites that include lots of recipes and make recipes easier to access than before. However, there is often the case that we could not get all the ingredients listed on the recipe. Prior research on alternative ingredient substitution has built a recommendation system considering the suitability of a recommended ingredient with the remained ingredients. In this paper, in addition to suitability, we also take the diversity of the ingredient categories and the novelty of new combination of ingredients into account. Besides, we combine suitability with novelty as an index, to see whether our method could help find out a new combination of ingredients that is possibly to be a new dish. Our evaluation results show that our proposed method attains a comparable or even better performance on each perspective.
Liu, Kuan-Hung; Chen, Hung-Chih; Lai, Kuan-Ting; Wu, Yi-Ying; and Wei, Chih-Ping, "Alternative Ingredient Recommendation: A Co-occurrence and Ingredient Category Importance Based Approach" (2018). PACIS 2018 Proceedings. 298.