Using the EM algorithm to train neural networks: misconceptions and a new algorithm for multiclass classification.
Shu-Kay NgGeoffrey J. McLachlanPublished in: IEEE Trans. Neural Networks (2004)
Keyphrases
- em algorithm
- expectation maximization
- multiclass classification
- neural network
- parameter estimation
- maximum likelihood
- gaussian mixture
- learning algorithm
- maximum likelihood estimation
- k means
- probabilistic model
- multi class
- mixture model
- generative model
- incomplete data
- objective function
- classification algorithm
- unsupervised learning
- training data
- image segmentation
- lower bound