A Data-Based Augmented Model Identification Method for Linear Errors-in-Variables Systems Based on EM Algorithm.
Fan GuoOuyang WuYongsheng DingBiao HuangPublished in: IEEE Trans. Ind. Electron. (2017)
Keyphrases
- em algorithm
- hidden variables
- parameter estimation
- expectation maximization
- incomplete data
- input data
- maximum likelihood estimation
- log likelihood
- mixture model
- gaussian mixture
- gibbs sampling
- likelihood function
- prior knowledge
- statistical methods
- maximum likelihood
- prior information
- probabilistic model
- missing data
- statistical model
- expectation maximisation
- hyperparameters
- generative model
- mixture components
- linear model
- gaussian mixture model
- parameter learning
- generative topographic mapping
- finite mixture models
- objective function
- probability density function
- closed form
- linear models
- mixture of gaussians
- model based clustering
- mixture distribution
- finite mixture model
- gaussian distribution
- linear regression
- maximum a posteriori
- energy function
- training data
- density estimation
- mixture modeling
- data points
- pairwise
- penalized likelihood
- estimate the model parameters
- probabilistic mixture model