Learning Partial Lexicographic Preference Trees and Forests over Multi-Valued Attributes

15 pages•Published: September 29, 2016

Abstract

\tit{Partial lexicographic preference trees}, or \tit{PLP-trees}, form an intuitive formalism for compact representation of qualitative preferences over combinatorial domains. We show that PLP-trees can be used to accurately model preferences arising in practical situations, and that high-accuracy PLP-trees can be effectively learned. We also propose and study learning methods for a variant of our model based on the concept of a PLP-forest, a collection of PLP-trees, where the preference order specified by a PLP-forest is obtained by aggregating the orders of its constituent PLP-trees. Our results demonstrate the potential of both approaches, with learning PLP-forests showing particularly promising behavior.

Keyphrases: learning preference models, partial lexicographic preference forests, partial lexicographic preference trees, preference reasoning, preference representation

In: Christoph Benzmüller, Geoff Sutcliffe and Raul Rojas (editors). GCAI 2016. 2nd Global Conference on Artificial Intelligence, vol 41, pages 314-328.

Links:	https://easychair.org/publications/paper/6z
	https://doi.org/10.29007/xtl4