Corpus
Question 1: What is a corpus?
A large set of structured text that has been produced in a natural communicative setting
Question 2: What is representativeness?
It is a measure of how well learning a corpus can generalize to the entire language
Question 3: What are the two factors in determining the representativeness of a corpus?
- Balance -> No. of genres to chose from
- Chunking -> How the chunk for each genre is chosen
Question 4: What is a tree bank
It is a corpus that has been linguistically parsed and annotated with semantic/syntactic structure
Question 5: What are the two types of treebank corpus
- Semantic
- Syntactic
Question 6: What to semantic treebanks use?
Formal representation of the semantic structure
Question 7: Give an example to semantic treebank
- Robocup
- Robot Commands
Question 8: What do syntactic treebanks use?
The use predicate logic based meaning representation
Question 9: Give an example to syntactic treebank
- Penn arabic
- Columbia arabic
Question 10: What are the application of treebanks
- Computational Linguistics to build state-of-the-art NLP system to do POS tagging, parsing, analyzing etc…
- In Corpus Linguistics to study syntactic phenomena
- In Theoretical linguistics and psycholinguistics, interaction evidence
Question 11: What is a PropBank?
It is a corpus annotated with verbal prepositions and their arguments
Question 12: What is the use of PropBank?
It helps in semantic role labelling
Question 13: What is the distinguishing feature of WordNet?
Nouns, verbs, adverbs and adjectives are organized into groups known as SynSets They are connected using conceptual semantic and lexical relation