EM Algorithm is Sample-Optimal for Learning Mixtures of Well-Separated Gaussians.
Jeongyeol KwonConstantine CaramanisPublished in: CoRR (2020)
Keyphrases
- em algorithm
- mixture model
- expectation maximization
- gaussian mixture
- gaussian mixture model
- hidden variables
- generative model
- learning algorithm
- maximum likelihood
- density estimation
- maximum likelihood estimation
- parameter learning
- likelihood maximization
- probability density function
- parameter estimation
- unsupervised learning
- likelihood function
- model based clustering
- mixture modeling
- reinforcement learning
- closed form
- multi modal
- incomplete data
- state space
- prior knowledge
- high dimensional
- pairwise
- finite mixture models