Project Outline and Aims
Classical recommender systems provide users with a list of recommendations where each recommendation consists of a single item, e.g., a book or DVD. However, applications such as travel planning can benefit from a system capable of recommending packages of items in the form of sets or sequences, and within a user-specified budget. In this context, it is useful if the system can recommend the top-k packages for the user to choose from. Motivated by these applications, we are studying composite recommendations, where each recommendation comprises a set or sequence of items. Each item has both a value (rating) and a cost associated with it, and the user specifies a maximum total cost (budget) for any composite recommendation. Because the problem of generating the top recommendation (package) is NP-complete, we are studying approximation algorithms for generating top-k composite recommendations.
Funding and Staffing Details
- M. Xie, L.V.S. Lakshmanan and P.T. Wood, ``Efficient Rank Join with Aggregation Constraints,'' in Proc. 37th Int. Conf. on Very Large Data Bases (Aug. 29-Sept. 3, Seattle, Washington), 2011.
- M. Xie, L.V.S. Lakshmanan and P.T. Wood, ``CompRec-Trip: a Composite Recommendation System for Travel Planning,'' demo paper in Proc. 27th IEEE Int. Conf. on Data Engineering (Hannover, Germany, Apr. 11-16), 2011.
- M. Xie, L.V.S. Lakshmanan and P.T. Wood, ``Breaking out of the Box of Recommendations: From Items to Packages,'' to appear in 4th ACM Conf. on Recommender Systems (Barcelona, Spain, Sept. 26-30), 2010.