Any reasonable cost function can be used for a posteriori probability approximation.
Marco SaerensPatrice LatinneChristine DecaesteckerPublished in: IEEE Trans. Neural Networks (2002)
Keyphrases
- cost function
- probability distribution
- error bounds
- central limit theorem
- global minimum
- maximum likelihood
- approximation algorithms
- relative error
- confidence level
- approximation error
- approximation methods
- high quality
- closed form
- conditional probabilities
- image processing
- expected cost
- support vector
- case study
- information systems