Collective Multi-Label Classification
Abstract
Common approaches to multi-label classification learn independent classifiers for each category, and employ ranking or thresholding schemes for classification. Because they do not exploit dependencies between labels, such techniques are only well-suited to problems in which categories are independent. However, in many domains labels are highly interdependent. This paper explores multi-label conditional random field (CRF) classification models that directly parameterize label co-occurrences in multi-label classification. Experiments show that the models outperform their single-label counterparts on standard text corpora. Even when multi-labels are sparse, the models improve subset classification error by as much as 40%.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 01, 2005
- Accession Number
- ADA440081
Entities
People
- Andrew McCallum
- Nadia Ghamrawi
Organizations
- University of Massachusetts Amherst