Multi-Instance Classification by Max-Margin Training of Cardinality-Based Markov Networks.
Hossein HajimirsadeghiGreg MoriPublished in: IEEE Trans. Pattern Anal. Mach. Intell. (2017)
Keyphrases
- markov networks
- max margin
- structured prediction
- document classification
- maximum margin
- graphical models
- first order logic
- maximum likelihood
- multi label
- multi label classification
- training set
- belief propagation
- conditional random fields
- bayesian networks
- support vector
- support vector machine
- probabilistic model
- supervised learning
- markov random field
- image classification
- training samples
- bayesian inference
- text classification
- classification accuracy
- feature vectors
- decision trees
- inductive logic programming
- hidden variables
- classification algorithm
- multi class
- posterior probability
- text categorization
- conditional probabilities
- latent variables
- class labels
- machine learning
- feature extraction
- feature selection
- hyperplane
- training examples
- model selection
- feature space