Multi-prototype Label Ranking with Novel Pairwise to Total Rank Aggregation

Publication
Aug 6, 2013
Abstract

We propose a Multi-prototype-based algorithm for online learning of soft pairwise-preferences over labels. The algorithm learns the label preferences via minimization of the proposed soft rank-loss measure, and can learn from total orders as well as from various types of partial orders. The soft pairwise preference algorithm outputs are further aggregated to produce a total label ranking prediction using a novel aggregation algorithm that outperforms ex- isting aggregation solutions. Experiments on synthetic and real-world data demonstrate state-of-the- art performance of our model.

  • International Joint Conference on Artificial Intelligence (IJCAI)
  • Conference/Workshop Paper

BibTeX