A synthetic approach to Markov kernels, conditional independence, and theorems on sufficient statistics.
Tobias FritzPublished in: CoRR (2019)
Keyphrases
- conditional independence
- sufficient statistics
- bayesian networks
- hidden variables
- graphical models
- data points
- random variables
- causal models
- chain graphs
- structure learning
- k means
- directed acyclic graph
- probability distribution
- axiomatic characterization
- markov property
- support vector
- kernel function
- transition probabilities
- kernel methods
- expectation maximization
- probabilistic model
- latent variables
- clustering algorithm
- lower bound
- text classification
- active learning