Abstract
Data-driven parsers rely on recommendations from parse models, which are generated from a set of training data using a machine learning classifier, to perform parse operations. However, unless the training data covers every possible situation there may be cases where a parse model is unable to make a recommendation. Therefore, when a parse model recommends no/several parse actions to a parser, it will be hard for a parser to make an informed decision as to what parse operation to perform. Here we examine the effect of various deterministic choices on a data-driven parser when it is presented with no/several recommendation from a parse model.
Original language | English |
---|---|
Title of host publication | The 12th International Workshop on Natural Language Processing and Cognitive Science |
Place of Publication | Poland |
Publication status | Published - 22 Sept 2015 |
Event | The 12th International Workshop on Natural Language Processing and Cognitive Science - Jagiellonian University Duration: 22 Sept 2015 → 24 Sept 2015 |
Conference
Conference | The 12th International Workshop on Natural Language Processing and Cognitive Science |
---|---|
City | Jagiellonian University |
Period | 22/09/15 → 24/09/15 |
Keywords
- Data-driven parsing
- Deterministic parsing
- Natural language parsing
- Arabic parsing