Activity Number:
|
131
- Predictive Modeling in Data Science
|
Type:
|
Contributed
|
Date/Time:
|
Monday, July 31, 2017 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Section on Statistical Learning and Data Science
|
Abstract #323099
|
|
Title:
|
An Overview of Existing and a Novel Approaches to Multi-Label Classification
|
Author(s):
|
Hyukjun Gweon* and Matthias Schonlau and Stefan Steiner
|
Companies:
|
and University of Waterloo and University of Waterloo
|
Keywords:
|
Multi-label classification ;
Label correlation ;
Machine learning
|
Abstract:
|
Multi-label classification is a supervised learning problem where an observation may be associated with multiple binary (outcome) labels simultaneously. We give an overview over common approaches to multi-label classification and also introduce a new approach as an extension of the nearest neighbor principle. Experiments on benchmark multi-label data sets show that the proposed method on average outperforms other commonly used approaches in terms of classification performance.
|
Authors who are presenting talks have a * after their name.