Bootstrapping a Tagalog LFG F-structure Bank

Developing a wide coverage grammar is an expensive and time-consuming process. Automatic induction methods use treebanks to produce F-structures which will then be used to induce the grammar. However, this method of automatic induction for creating a Filipino LFG grammar is not possible because curr...

Full description

Saved in:
Bibliographic Details
Main Authors: Cada, Djesus Rex C., Chan, Francis Alexis L., Chen, Hua Zong S., Tan, Andrew Edwell Y.
Format: text
Language:English
Published: Animo Repository 2012
Online Access:https://animorepository.dlsu.edu.ph/etd_bachelors/11933
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: De La Salle University
Language: English
Description
Summary:Developing a wide coverage grammar is an expensive and time-consuming process. Automatic induction methods use treebanks to produce F-structures which will then be used to induce the grammar. However, this method of automatic induction for creating a Filipino LFG grammar is not possible because currently, there are no existing treebanks for Filipino. Bootstrapping was used to solve the aforesaid problem. From existing parallel English-Filipino corpora, cross-lingual projection was used to project F-structure from English. Using state of the art and self-developed tools, equivalent Tagalog words from the parallel corpora were mapped into the projected F-structures to produce Tagalog F-structures. The final output of the research is a collection of Tagalog F-structures induced from English F-structure.