Decreasing Lexical Data Sparsity in Statistical Syntactic Parsing - Experiments with Named Entities
Year: 2011
Type: conference-paper
Abstract: In this paper we present preliminary experiments that aim to reduce lexical data sparsity in statistical parsing by exploiting information about named entities. Words in the\nWSJ corpus are mapped to named entity clusters and a latent variable constituency parser is trained and tested on the transformed corpus. We explore two different methods for\nmapping words to entities, and look at the effect of mapping various subsets of named entity types. Thus far, results show no improvement in parsing accuracy over the best baseline score; we identify possible problems and outline suggestions for future directions.
Institution: Dublin City University
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W200741329