I have annotated my dataset by using POS, chunking, and wordcase. If I include dependency parser in my annotated dataset, will it help to define more features for the classification of Movie Name Entity? In short, will dependency relations improve the performance of the model for Move Reviews dataset? I need to identify movie names and person names from my corpus.
there is no link associated with your answer. Please complete the answer!
If you you've defined POS tags, you've surely come across proper nouns that can define Name Entities. Regarding features: The answer is: maybe. Out of my personal experience: the more features, the better results. However, don't forget the data overfitting issue. Regarding dependencies: It depends on what you understand under dependencies. There is a Stanford dependency parser, but it not clear how it works since the documentation is very vague.
- Badges
- Science topic
- Similar topics
- Biological Science
- Anatomy
- Tissue