CookRecipe: A Recipe Analysis and Learning System

The recent explosive proliferation of interesting and useful data over the Web such as various recipes, while providing people with readily available information, brings out a challenging issue on how to manage such non-conventional data effectively. To respond the challenge, the researchers propose to develop a Web-based recipe database system to manage recipes in a novel way, which not only covers the static recipe attributes but also elucidates the dynamic cooking behaviors. In this project, the researchers shall devise several semantic knowledge facilities, with which we lay down a solid foundation for the system to provide users with such versatile services as personalized adaptation and recipe recommendation. More specifically, they will work on the fundamentals of the proposed system, particularly the development of a recipe data model embodying a cooking graph, associated constraints and some properties inherent to a recipe. The researchers will also study how to discover cooking techniques by mining patterns from real recipe data; similarity calculation based on cooking graphs will be investigated, and a query-by-example recipe search mechanism is to be devised (especially when the example recipe has been modified by the user in advance). The usefulness and elegancy of this recipe database system will be demonstrated through building a prototype system, upon which efficiency issues of the underlying algorithms will be addressed through experimental studies.