Abstract
Conditional preference networks (CP-nets) provide a compact and intuitive graphical tool to represent the preferences of a user. However, learning such a structure is known to be a difficult problem due to its combinatorial nature. We propose, in this paper, a new, efficient, and robust query-based learning algorithm for acyclic CP-nets. In particular, our algorithm takes into account the contradictions between multiple users’ preferences by searching in a principled way the variables that affect the preferences. We provide complexity results of the algorithm, and demonstrate its efficiency through an empirical evaluation on synthetic and on real databases.
Similar content being viewed by others
Notes
An object can be a hotel having as a set of attributes: the number of rooms, the price, etc.
A partial ordering \(\succ\) is an asymmetric, irreflexive, and transitive relation, i.e., if \(\mathbf {x} \succ \mathbf {y}\), then \(\mathbf {y} \not \succ \mathbf {x}\) (asymmetry), \(\mathbf {x} \not \succ \mathbf {x}\) (irreflexivity), and if \(\mathbf {x} \succ \mathbf {y}\) and \(\mathbf {y} \succ \mathbf {z}\), then \(\mathbf {x} \succ \mathbf {z}\) (transitivity).
A preference model \(\mathcal {P}\) corresponds to a model where the preferences have properties, such that conditional preferences, additive preferences, etc.
We say that two preference models \(\mathcal {P}\) and \(\mathcal {P}'\) are equivalent, denoted by \(\mathcal {P} \equiv \mathcal {P}^\prime\) iff they induce exactly the same preferences.
We consider a variable V as confident if enough swaps that induce the rules of V are found.
Two outcomes \(\mathbf {o}\) and \(\mathbf {o}'\) are comparable if either \(\mathbf {o} \succ \mathbf {o}'\) or \(\mathbf {o}' \succ \mathbf {o}\).
We say that a CP-net is sparse if the adjacency matrix of its corresponding graph is sparse.
References
Alanazi E, Mouhoub M, Zilles S (2016) The complexity of learning acyclic cp-nets. In: Proceedings of the Twenty-Fifth International Joint Conference on Artificial Intelligence, IJCAI 2016, New York, NY, USA, 9–15 July 2016, pp 1361–1367. http://www.ijcai.org/Abstract/16/196
Angluin D (1987) Queries and concept learning. Mach Learn 2(4):319–342. doi:10.1007/BF00116828
Boutilier C, Brafman RI, Domshlak C, Hoos HH, Poole D (2004) Cp-nets: a tool for representing and reasoning with conditional ceteris paribus preference statements. J Artif Intell Res 21:135–191. doi:10.1613/jair.1234
Chevaleyre Y, Koriche F, Lang J, Mengin J, Zanuttini B (2010) Learning ordinal preferences on multiattribute domains: the case of cp-nets. In: Preference Learning, pp 273–296. doi:10.1007/978-3-642-14125-6_13
Dimopoulos Y, Michael L, Athienitou F (2009) Ceteris paribus preference elicitation with predictive guarantees. In: Proceedings of the 21st International Joint Conference on Artificial Intelligence, IJCAI 2009, Pasadena, California, USA, 11–17 July 2009, pp 1890–1895. http://ijcai.org/Proceedings/09/Papers/313.pdf
Eckhardt A, Vojtás P (2009) How to learn fuzzy user preferences with variable objectives. In: Proceedings of the Joint 2009 International Fuzzy Systems Association World Congress and 2009 European Society of Fuzzy Logic and Technology Conference, Lisbon, Portugal, 20–24 July 2009, pp 938–943. http://www.eusflat.org/proceedings/IFSA-EUSFLAT_2009/pdf/tema_0938.pdf
Eckhardt A, Vojtás P (2010) Learning user preferences for 2cp-regression for a recommender system. In: SOFSEM 2010: Theory and Practice of Computer Science, 36th Conference on Current Trends in Theory and Practice of Computer Science, Spindleruv Mlýn, Czech Republic, 23–29 January 2010, pp 346–357. doi:10.1007/978-3-642-11266-9_29
Fürnkranz J, Hüllermeier E (eds) (2010) Preference Learning. Springer, Berlin. doi:10.1007/978-3-642-14125-6
Guerin JT, Allen TE, Goldsmith J (2013) Learning cp-net preferences online from user queries. In: Late-Breaking Developments in the Field of Artificial Intelligence, Bellevue,Washington, USA, 14–18 July 2013. http://www.aaai.org/ocs/index.php/WS/AAAIW13/paper/view/7114
Koriche F, Zanuttini B (2010) Learning conditional preference networks. Artif Intell 174(11):685–703. doi:10.1016/j.artint.2010.04.019
Liu J, Yao Z, Xiong Y, LiuW,Wu C (2013) Learning conditional preference network from noisy samples using hypothesis testing. Knowl Based Syst 40:7–16. doi:10.1016/j.knosys.2012.11.006
Liu J, Xiong Y, Wu C, Yao Z, Liu W (2014) Learning conditional preference networks from inconsistent examples. IEEE Trans Knowl Data Eng 26(2):376–390. doi:10.1109/TKDE.2012.231
Liu J, Sui C, Deng D, Wang J, Feng B, Liu W, Wu C (2016) Representing conditional preference by boosted regression trees for recommendation. Inf Sci 327:1–20. doi:10.1016/j.ins.2015.08.001
Michael L, Papageorgiou E (2013) An empirical investigation of ceteris paribus learnability. In: IJCAI 2013, Proceedings of the 23rd International Joint Conference on Artificial Intelligence, Beijing, China, 3–9 August 2013, pp 1537–1543. http://www.aaai.org/ocs/index.php/IJCAI/IJCAI13/paper/view/6923
Tsoukiàs A (2008) From decision theory to decision aiding methodology. Eur J Operat Res 187(1):138–161. doi:10.1016/j.ejor.2007.02.039
Wang H, Lu Y, Zhai C (2010) Latent aspect rating analysis on review text data: a rating regression approach. In: Proceedings of the 16th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Washington, DC, USA, 25–28 July 2010, pp 783–792. doi:10.1145/1835804.1835903
Wang H, Lu Y, Zhai C (2011) Latent aspect rating analysis without aspect keyword supervision. In: Proceedings of the 17th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Diego, CA, USA, 21–24 August 2011, pp 618–626. doi:10.1145/2020408.2020505
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Labernia, F., Yger, F., Mayag, B. et al. Query-based learning of acyclic conditional preference networks from contradictory preferences. EURO J Decis Process 6, 39–59 (2018). https://doi.org/10.1007/s40070-017-0070-3
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s40070-017-0070-3