fined in a workflow for better performance. One may go further [12] D. Goldberg, D. Nichols, B Oki, and D. Terry. Using and extend the database query engine with the processing capabili ollaborative filtering to weave an information tapestry. C. of ties required for supporting recommendations. M,35(12):61-70.1992 In addition, FlexRecs define recommendations over relational [13] J Herlocker, J Konstan, L. Terveen, and J. Riedl. Evaluating data. Large amounts of data are relational. Our university data ollaborative filtering recommender systems TO/S, 22: 5-53 are also relational. It would be interesting to define flexible recom 2004. endations for XML or ontologies. For example, XML is a more [14 G Karypis. Evaluation of item-based top-n recommendation flexible data model for handling nested relations but there are chal- algorithms In CIKM, 200 lenges in defining and implementing the recommendation oper [15] G. Koutrika, R Ikeda, B. Bercovitz, and H. Garcia-Molina tors. Finally, our FlexRecs system serves as an extensible research Flexible recommendations over rich data. In RecSys, pages and development platform for novel recommendation workflows. 203-210,2008. le are experimenting with different recommendation types, such [16 G. Koutrika, B. Bercovitz, R Ikeda, F. Kaliszan, H. Liou, as recommending majors or blending methods. Part of this effort and H. Garcia-Molina. Flexible Recommendations for is designing appropriate user interfaces for enabling users express Course planning In /CDE. 2009 fexible reco [17G.LindensB.SmithandJ.York.Amazon.com 7. REFERENCES commendations: Item-to-item collaborative filtering. IEEE [1 Problems with recommendations: url Internet Computing, Jan/Feb 2003 http://www.amazon.com/problemswith-kindle-store [18] M. Mahony, N. Hurley. N. Kushmerick, and G Silvestre. recommendations/forum/fxbvkstO6pwp 9b/txltogp7gq4xa4t Collaborative recommendation: A robustness analysis. ACM =utf8&asin=b0o0fi73ma TOT,4(4):344-377,2004. [2] The Stanford Daily: url: [19 B. Miller, I. Albert, S. Lam, J. Konstan, and J. Riedl. http://stanforddaily.com/article/2007/1 Movielens unplugged: Experiences with an occasional ditorialcourserankalongoverduesuccess onnected recommender system. In Int'l Conf. Intelli 3] G. Adomavicius and Y Kwon. New recommendation ser Interfaces, 2003 techniques for multi-criteria rating systems. IEEE Intelligent 120)B Mobasher, R. Burke, RBhaumik, and C Williams. Systems,22(3),2007 Towards trustworthy recommender systems: An analysis of 4 G. Adomavicius and A. Tuzhilin Multidimensional attack models and algorithm robustness. ACM Trans. or recommender systems: A data warehousing approach. In Internet Technology, 7(2), 2007 WELCOM. 2001 [211 M. Pazzani and D Billsus. Learning and revising user 5] G. Adomavicius and A Tuzhilin rofiles. The identification of interesting web sites machine of recommender systems: A surve armIng,27:313-331,1997 and possible extensions. IEEE Tr Knowledge [22] P Resnick, N. lakovou, M. Sushak, P Bergstrom, and Data Engineering, 17(6): 734-749, 2005 J. Riedl. Grouplens: An open architecture for collaborative [6 M. Balabanovic and Y Shoham Fab filtering of netnews. In Conf on Computer Supported collaborative recommendation. C of ACM, 40(3): 66-72, Cooperative Work, 1994. [23] B Sarwar, G. Karypis, J. Konstan, and J Riedl. Item-based [7 D. Billsus and M. Pazzani. User modeling for adaptive news collaborative filtering recommendation algorithms. In 10 access User Modeling and User-Adapted interaction www Conf, 2001 10(2-3):147-180,2000 [24]A Schein, A Popescul, L. Ungar, and D. Pennock. Methods [8] J Breese, D Heckerman, and C. Kadie Empirical analysis and metrics for cold-start recommendations In ACM SIGIR f predictive algorithms for collaborative filtering. In 14th Conf.,2002 [25] H J Schek and M. H Scholl. The relational model with [9] L. Colby. A recursive algebra for nested relations. Inf. Sy relation-valued attributes. Inf. Syst, 11(2): 137-147, 1986 15(5):567-582,1990. [26] U Shardanand and P Maes. Social information filtering [10 A Das, M. Datar, A Garg, and S Rajaram. Google news Algorithms for automating word of mouth. In Confon ersonalization: scalable online collaborative filtering. In Human Factors in Comp. Sys., 1995 www, pages271-280,2007 [27 B. Sheth and P. Maes Evolving agents for personalized [11 V. Deshpande and P. Larson. Transforming from flat algebra information filtering In JEE Artificial Intelligence for to nested algebra System Sciences, 2(2-5): 298-307, 1990 Applications, 1993defined in a workflow for better performance. One may go further and extend the database query engine with the processing capabili￾ties required for supporting recommendations. In addition, FlexRecs define recommendations over relational data. Large amounts of data are relational. Our university data are also relational. It would be interesting to define flexible recom￾mendations for XML or ontologies. For example, XML is a more flexible data model for handling nested relations but there are chal￾lenges in defining and implementing the recommendation opera￾tors. Finally, our FlexRecs system serves as an extensible research and development platform for novel recommendation workflows. We are experimenting with different recommendation types, such as recommending majors or blending methods. Part of this effort is designing appropriate user interfaces for enabling users express flexible recommendations. 7. REFERENCES [1] Problems with recommendations: url: http://www.amazon.com/problemswith-kindle-store￾recommendations/forum/fxbvkst06pwp9b/tx1togp7gq4xa4t /1?_encoding=utf8&asin=b000fi73ma. [2] The Stanford Daily: url: http://stanforddaily.com/article/2007/12/5/- editorialcourserankalongoverduesuccess. [3] G. Adomavicius and Y. Kwon. New recommendation techniques for multi-criteria rating systems. IEEE Intelligent Systems, 22(3), 2007. [4] G. Adomavicius and A. Tuzhilin. Multidimensional recommender systems: A data warehousing approach. In WELCOM, 2001. [5] G. Adomavicius and A. Tuzhilin. Toward the next generation of recommender systems: A survey of the state-of-the-art and possible extensions. IEEE Trans. on Knowledge and Data Engineering, 17(6):734–749, 2005. [6] M. Balabanovic and Y. Shoham. Fab: Content-based, collaborative recommendation. C. of ACM, 40(3):66–72, 1997. [7] D. Billsus and M. Pazzani. User modeling for adaptive news access. User Modeling and User-Adapted Interaction, 10(2-3):147–180, 2000. [8] J. Breese, D. Heckerman, and C. Kadie. Empirical analysis of predictive algorithms for collaborative filtering. In 14th UAI Conf., 1998. [9] L. Colby. A recursive algebra for nested relations. Inf. Syst., 15(5):567–582, 1990. [10] A. Das, M. Datar, A. Garg, and S. Rajaram. Google news personalization: scalable online collaborative filtering. In WWW, pages 271–280, 2007. [11] V. Deshpande and P. Larson. Transforming from flat algebra to nested algebra. System Sciences, 2(2-5):298–307, 1990. [12] D. Goldberg, D. Nichols, B. Oki, and D. Terry. Using collaborative filtering to weave an information tapestry. C. of ACM, 35(12):61–70, 1992. [13] J. Herlocker, J. Konstan, L. Terveen, and J. Riedl. Evaluating collaborative filtering recommender systems. TOIS, 22:5–53, 2004. [14] G. Karypis. Evaluation of item-based top-n recommendation algorithms. In CIKM, 2001. [15] G. Koutrika, R. Ikeda, B. Bercovitz, and H. Garcia-Molina. Flexible recommendations over rich data. In RecSys, pages 203–210, 2008. [16] G. Koutrika, B. Bercovitz, R. Ikeda, F. Kaliszan, H. Liou, and H. Garcia-Molina. Flexible Recommendations for Course Planning. In ICDE, 2009. [17] G. Linden, B. Smith, and J. York. Amazon.com recommendations: Item-to-item collaborative filtering. IEEE Internet Computing, Jan/Feb 2003. [18] M. Mahony, N. Hurley, N. Kushmerick, and G. Silvestre. Collaborative recommendation: A robustness analysis. ACM TOIT, 4(4):344–377, 2004. [19] B. Miller, I. Albert, S. Lam, J. Konstan, and J. Riedl. Movielens unplugged: Experiences with an occasionally connected recommender system. In Int’l Conf. Intelligent User Interfaces, 2003. [20] B. Mobasher, R. Burke, R. Bhaumik, and C. Williams. Towards trustworthy recommender systems: An analysis of attack models and algorithm robustness. ACM Trans. on Internet Technology, 7(2), 2007. [21] M. Pazzani and D. Billsus. Learning and revising user profiles: The identification of interesting web sites. Machine Learning, 27:313–331, 1997. [22] P. Resnick, N. Iakovou, M. Sushak, P. Bergstrom, and J. Riedl. Grouplens: An open architecture for collaborative filtering of netnews. In Conf. on Computer Supported Cooperative Work, 1994. [23] B. Sarwar, G. Karypis, J. Konstan, and J. Riedl. Item-based collaborative filtering recommendation algorithms. In 10th WWW Conf., 2001. [24] A. Schein, A. Popescul, L. Ungar, and D. Pennock. Methods and metrics for cold-start recommendations. In ACM SIGIR Conf., 2002. [25] H. J. Schek and M. H. Scholl. The relational model with relation-valued attributes. Inf. Syst., 11(2):137–147, 1986. [26] U. Shardanand and P. Maes. Social information filtering: Algorithms for automating ‘word of mouth’. In Conf. on Human Factors in Comp. Sys., 1995. [27] B. Sheth and P. Maes. Evolving agents for personalized information filtering. In IEEE Conf. Artificial Intelligence for Applications, 1993