[OPENNLP-493] TokenNameFinderTrainer should have an option to only use certain name types for training

• Type: Improvement
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: tools-1.5.3
• Component/s: Command Line Interface, Name Finder
• Assignee: Jörn Kottmann
• Reporter: Jörn Kottmann
• Created: 04/Apr/12 13:05
• Updated: 27/Apr/12 11:10
• Resolved: 27/Apr/12 10:54

Description

Training data produced from a corpus (e.g. conll, muc) often contain many name types, but usually the training should just be done for one name type (or a few selected ones). To make this easier the name finder trainer should have an optional argument to specify the types to use for training.