A Chinese dishes recommendation algorithm based on personal taste

Abstract

There are few Chinese dish recommendation algorithms due to the variety of Chinese dishes. It could be impossible to find one's most liked dishes in a restaurant through the name or the ingredients of a dish. The algorithm in this paper uses the user's ordering history to quantify one's taste by k-means clustering method and determines the number of user's favorite tastes by the BWP index. With the knowledge of user's tastes, screen matrix are used to rank the dishes according to the user's taste in any restaurant.