Label Ranking with Abstention: Predicting Partial Orders by Thresholding Probability Distributions

author: Weiwei Cheng, Mathematik und Informatik, Philipps-Universität Marburg
published: Jan. 24, 2012, recorded: December 2011, views: 217

Slides

Slides
0:00	Label Ranking with Abstention
1:04	Label Ranking – An Example -1
1:38	Label Ranking – An Example -2
2:05	Label Ranking
2:26	Learning with Reject Option
3:03	From Total to Partial Order Relations
4:14	Partial Orders from Pairwise Comparisons -1
5:46	Partial Orders from Pairwise Comparisons -2
6:00	Partial Orders from Pairwise Comparisons -3
6:26	Partial Orders from Pairwise Comparisons -4
7:14	Our Ideas & Results
8:10	The Plackett-Luce Model -1
9:15	The Plackett-Luce Model -2
9:42	The Plackett-Luce Model -3
9:52	The Plackett-Luce Model -4
10:02	The Plackett-Luce Model -5
10:16	The Mallows Model
11:45	Some Common Choices of Δ
12:34	Transposition Property
14:08	Remarks on P (yi > yj)
15:48	Our Main Result
16:50	Experimental Results
18:37	Take-Away Messages
19:15	Thank You
20:24	Remarks on P (yi > yj)

Report a problem or upload files

If you have found a problem with this lecture or would like to send us extra material, articles, exercises, etc., please use our ticket system to describe your request and upload the data.
Enter your e-mail into the 'Cc' field, and we will keep you updated with your request's status.

Lecture popularity: You need to login to cast your vote.

Description

We consider an extension of the setting of label ranking, in which the learner is allowed to make predictions in the form of partial instead of total orders. Predictions of that kind are interpreted as a partial abstention: If the learner is not sufficiently certain regarding the relative order of two alternatives, it may abstain from this decision and instead declare these alternatives as being incomparable. We propose a new method for learning to predict partial orders that improves on an existing approach, both theoretically and empirically. Our method is based on the idea of thresholding the probabilities of pairwise preferences between labels as induced by a predicted (parameterized) probability distribution on the set of all rankings.