Abstract
The naive Bayes assumption in text classification has the advantage of greatly simplifying maximum likelihood estimation of unknown class-conditional word occurrence probabilities. However, these estimates are usually modified by application of a heuristic parameter smoothing technique to avoid (over-fitted) null estimates. In this work, we advocate the reduction of the parameter domain instead of parameter smoothing. This leads to a constrained domain maximum likelihood estimation problem for which we provide an iterative algorithm that solves it optimally.
Similar content being viewed by others
References
Boyd S, Vandenberghe L (2004) Convex optimization. Cambridge University Press, London
Hoare Z (2007) Landscapes of naïve bayes classifiers. Pattern Anal Appl 11(1):59–72
Juan A, Ney H (2002) Reversing and smoothing the multinomial naive Bayes text classifier. In Proceedings of PRIS 2002, pp 200–212
Juan A, Vilar D, Ney H (2007) Bridging the gap between Naive Bayes and maximum entropy. In Proceedings of the PRIS 2007, Funchal (Portugal), pp 59–65
Lewis DD (1998) Naive Bayes at forty: the independence assumption in information retrieval. In: Proceedings of the ECML’98, pp 4–15
McCallum A (1998) Bow: a toolkit for statistical language modeling, text retrieval, classification and clustering. http://www.cs.umass.edu/∼mccallum/bow/rainbow
McCallum A (2002) Industry sector data set. http://www.cs.umass.edu/∼mccallum/code-data.html
McCallum A, Nigam K (1998) A comparison of event models for Naive Bayes text classification. In: Proceedings of AAAI/ICML-98 workshop on learning for text categorization, pp 41–48
Rennie J (2001) Original 20 newsgroups data set, people.csail.mit.edu/ jrennie/ 20Newsgroups
Vidal E et al (2000) Final report of ESPRIT Research Project 30268 (EuTrans)
Vilar D, Ney H, Juan A, Vidal E (2004) Effect of feature smoothing methods in text classification tasks. In: Proceedings of PRIS 2004, pp 108–117
Acknowledgements
Work partially supported by the Spanish research programme Consolider Ingenio 2010: MIPRCV (CSD2007-00018), by the EC (FEDER), the Spanish MEC under grant TIN2006-15694-CO2-01 and the Valencian “Conselleria d’Empresa, Universitat i Ciència” under grant CTBPRA/2005/004.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Andrés-Ferrer, J., Juan, A. Constrained domain maximum likelihood estimation for naive Bayes text classification. Pattern Anal Applic 13, 189–196 (2010). https://doi.org/10.1007/s10044-009-0149-y
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10044-009-0149-y