Universal Lower Bounds and Optimal Rates: Achieving Minimax Clustering Error in Sub-Exponential Mixture Models.
Maximilien DrevetonAlperen GözetenMatthias GrossglauserPatrick ThiranPublished in: COLT (2024)
Keyphrases
- mixture model
- lower bound
- expected error
- worst case
- mixture modeling
- unsupervised learning
- density estimation
- model based clustering
- gaussian mixture model
- em algorithm
- upper bound
- generative model
- overlapping clustering
- finite mixture models
- probabilistic model
- k means
- optimal solution
- bayesian information criterion
- clustering method
- model selection
- expectation maximization
- clustering algorithm
- finite mixture model
- maximum likelihood
- probability density function
- finite mixtures
- language model
- object recognition
- estimation error
- minimum message length
- automatic model selection
- hierarchical clustering
- generalization error
- outlier detection
- data points
- feature selection
- bayesian networks
- reinforcement learning
- objective function
- exponential family
- sample size