Multi-Prototype Label Ranking with Novel Pairwise-to-Total-Rank Aggregation
Abstract
We propose a multi-prototype-based algorithm for online learning of soft pairwise-preferences over labels. The algorithm learns soft 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 existing aggregation solutions. Experiments on synthetic and real-world data demonstrate state-of-the-art performance of the proposed model.
Cite
Text
Grbovic et al. "Multi-Prototype Label Ranking with Novel Pairwise-to-Total-Rank Aggregation." International Joint Conference on Artificial Intelligence, 2013.Markdown
[Grbovic et al. "Multi-Prototype Label Ranking with Novel Pairwise-to-Total-Rank Aggregation." International Joint Conference on Artificial Intelligence, 2013.](https://mlanthology.org/ijcai/2013/grbovic2013ijcai-multi/)BibTeX
@inproceedings{grbovic2013ijcai-multi,
title = {{Multi-Prototype Label Ranking with Novel Pairwise-to-Total-Rank Aggregation}},
author = {Grbovic, Mihajlo and Djuric, Nemanja and Vucetic, Slobodan},
booktitle = {International Joint Conference on Artificial Intelligence},
year = {2013},
pages = {1358-1364},
url = {https://mlanthology.org/ijcai/2013/grbovic2013ijcai-multi/}
}