ML Anthology

Semi-Supervised Learning of Mixture Models

Fábio Gagliardi Cozman, Ira Cohen, Marcelo Cesar Cirelo
ICML 2003 pp. 99-106
/icml/2003/cozman2003icml-semi/

Abstract

This paper analyzes the performance of semisupervised learning of mixture models. We show that unlabeled data can lead to an increase in classification error even in situations where additional labeled data would decrease classification error. We present a mathematical analysis of this "degradation" phenomenon and show that it is due to the fact that bias may be adversely affected by unlabeled data. We discuss the impact of these theoretical results to practical situations. ICML Proceedings of the Twentieth International Conference on Machine Learning

PDF ICML Semantic Scholar

Cite

Text

Cozman et al. "Semi-Supervised Learning of Mixture Models." International Conference on Machine Learning, 2003.

Markdown

[Cozman et al. "Semi-Supervised Learning of Mixture Models." International Conference on Machine Learning, 2003.](https://mlanthology.org/icml/2003/cozman2003icml-semi/)

BibTeX

@inproceedings{cozman2003icml-semi,
  title     = {{Semi-Supervised Learning of Mixture Models}},
  author    = {Cozman, Fábio Gagliardi and Cohen, Ira and Cirelo, Marcelo Cesar},
  booktitle = {International Conference on Machine Learning},
  year      = {2003},
  pages     = {99-106},
  url       = {https://mlanthology.org/icml/2003/cozman2003icml-semi/}
}