Voting-based ensemble learning for partial lexicographic preference forests over combinatorial domains

We study preference representation models based on partial lexicographic preference trees (PLP-trees). We propose to represent preference relations as forests of small PLP-trees (PLP-forests), and to use voting rules to aggregate orders represented by the individual trees into a single order to be taken as a model of the agent’s preference relation. We show that when learned from examples, PLP-forests have better accuracy than single PLP-trees. We also show that the choice of a voting rule does not have a major effect on the aggregated order, thus rendering the problem of selecting the “right” rule less critical. Next, for the proposed PLP-forest preference models, we develop methods to compute optimal and near-optimal outcomes, the tasks that appear difficult for some other common preference models. Lastly, we compare our models with those based on decision trees, which brings up questions for future research.

[1]  Lirong Xia,et al.  Sequential composition of voting rules in multi-issue domains , 2009, Math. Soc. Sci..

[2]  Miroslaw Truszczynski,et al.  Aggregating Conditionally Lexicographic Preferences Using Answer Set Programming Solvers , 2013, ADT.

[3]  A. Rapoport,et al.  An Empirical Evaluation of Six Voting Procedures: Do They Really Make Any Difference? , 1993, British Journal of Political Science.

[4]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.

[5]  Niall M. Fraser.  Ordinal preference representations , 1994 .

[6]  Miroslaw Truszczynski,et al.  Reasoning with Preference Trees over Combinatorial Domains , 2015, ADT.

[7]  Eyke Hüllermeier,et al.  Lexicographic preferences for predictive modeling of human decision making: A new machine learning method with an application in accounting , 2017, Eur. J. Oper. Res..

[8]  Nicholas Mattei,et al.  Empirical Evaluation of Voting Rules with Strictly Ordered Preference Data , 2011, ADT.

[9]  Nic Wilson,et al.  Efficient Inference and Computation of Optimal Alternatives for Preference Languages Based On Lexicographic Models , 2017, IJCAI.

[10]  Xudong Liu,et al.  Human-in-the-Loop Learning of Qualitative Preference Models , 2019, FLAIRS Conference.

[11]  Maria Luisa Bonet,et al.  A New Algorithm for Weighted Partial MaxSAT , 2010, AAAI.

[12]  Lirong Xia,et al.  Aggregating Conditionally Lexicographic Preferences on Multi-issue Domains , 2012, CP.

[13]  Jerome L. Myers,et al.  Research Design and Statistical Analysis , 1991 .

[14]  Barry O'Sullivan,et al.  Multi-language evaluation of exact solvers in graphical model discrete optimization , 2016, Constraints.

[15]  W. Gehrlein.  Condorcet's paradox and the likelihood of its occurrence: different perspectives on balanced preferences* , 2002 .

[16]  Miroslaw Truszczynski,et al.  Learning Partial Lexicographic Preference Trees over Combinatorial Domains , 2015, AAAI.

[17]  Miroslaw Truszczynski,et al.  Preference Learning and Optimization for Partial Lexicographic Preference Forests over Combinatorial Domains , 2018, FoIKS.

[18]  Yann Chevaleyre,et al.  Learning conditionally lexicographic preference relations , 2010, ECAI.

[19]  Mirek Truszczynski,et al.  Learning Partial Lexicographic Preference Trees and Forests over Multi-Valued Attributes , 2016, GCAI.

[20]  Nic Wilson.  Preference Inference Based on Lexicographic Models , 2014, ECAI.