Abstract
We consider the problems of aggregating ordinal data in the form of permutations under a number of constraints that arise in practical applications such as gene prioritization. To perform constrained aggregation under the constrained Kendall tau distance, we introduce an LP relaxation of the problem and prove that it provides a constant approximation solution. We proceed to describe a number of related methods that may also be used for solving constrained correlation clustering arising in community detection.
This is a joint work with Farzad Farnoud, Gregory Puleo and Fardad Raisali.