An Alternative Method of Training Probabilistic LR Parsers.
Mark-Jan NederhofGiorgio SattaPublished in: ACL (2004)
Keyphrases
- detection method
- preprocessing
- experimental evaluation
- training phase
- active learning
- significant improvement
- cost function
- training samples
- maximum likelihood
- computational cost
- training process
- high precision
- classification method
- synthetic data
- em algorithm
- principal component analysis
- support vector machine
- data sets
- prior knowledge
- pairwise
- computational complexity
- objective function
- bayesian networks