Preference Modeling with Context-Dependent Salient Features
This work addresses preference modeling for applications like recommendation systems or political districting, though it appears incremental as it builds on existing pairwise comparison frameworks.
The paper tackles the problem of estimating item rankings from noisy pairwise comparisons by addressing irrational choice patterns like intransitivity, proposing a salient feature preference model that explains systematic intransitivity and demonstrates strong performance on synthetic and real datasets.
We consider the problem of estimating a ranking on a set of items from noisy pairwise comparisons given item features. We address the fact that pairwise comparison data often reflects irrational choice, e.g. intransitivity. Our key observation is that two items compared in isolation from other items may be compared based on only a salient subset of features. Formalizing this framework, we propose the salient feature preference model and prove a finite sample complexity result for learning the parameters of our model and the underlying ranking with maximum likelihood estimation. We also provide empirical results that support our theoretical bounds and illustrate how our model explains systematic intransitivity. Finally we demonstrate strong performance of maximum likelihood estimation of our model on both synthetic data and two real data sets: the UT Zappos50K data set and comparison data about the compactness of legislative districts in the US.