Selection of likelihood parameters for complex models determines the effectiveness of Bayesian calibration.
Karl-Heinz RahnKlaus Butterbach-BahlChristian WernerPublished in: Ecol. Informatics (2011)
Keyphrases
- maximum likelihood
- parameter estimation
- complex systems
- parameter estimates
- free parameters
- bayesian networks
- statistical models
- parametric models
- physical systems
- prior distribution
- likelihood function
- dynamic bayesian networks
- likelihood model
- bayesian network models
- fine tuning
- prior probabilities
- markov chain monte carlo
- neural network
- posterior distribution
- camera calibration
- experimental data
- real world