inria-00548546, version 1
The Tradeoff Between Generative and Discriminative Classifiers
Guillaume Bouchard
a, 1Bill Triggs 2
16th IASC International Symposium on Computational Statistics (COMPSTAT '04) (2004) 721--728
Abstract: Given any generative classifier based on an inexact density model, we can define a discriminative counterpart that reduces its asymptotic error rate. We introduce a family of classifiers that interpolate the two approaches, thus providing a new way to compare them and giving an estimation procedure whose classification performance is well balanced between the bias of generative classifiers and the variance of discriminative ones. We show that an intermediate trade-off between the two strategies is often preferable, both theoretically and in experiments on real data.
- a – Xerox Research
- 1: Xerox Research Centre Europe (XRCE)
- Xerox
- 2: LEAR (IMAG-INRIA Rhône-Alpes / GRAVIR)
- CNRS : FR71 – CNRS : UMR5527 – INRIA – Université Joseph Fourier - Grenoble I – Institut National Polytechnique de Grenoble (INPG)
- Domain : Computer Science/Computer Vision and Pattern Recognition
- Keywords : Statistical computing – numerical algorithms
- inria-00548546, version 1
- http://hal.inria.fr/inria-00548546
- oai:hal.inria.fr:inria-00548546
- From: Team Lear
- Submitted for:
- Submitted on: Monday, 20 December 2010 09:09:34
- Updated on: Monday, 10 January 2011 10:12:33
Associated documents
PDF : 
See also
Export