A parsing algorithm for constituent structures of Tagalog

Tagalog is a free-word order language, thus, exhibiting the problem of discontinuous constituents. Current approaches to resolve this type of problem use the configurational (or hierarchical) syntactic structure as a model to represent surface ordering of constituents, which is suitable for fixed-wo...

Full description

Saved in:
Bibliographic Details
Main Author: Dimalen, Editha D.
Format: text
Language:English
Published: Animo Repository 2003
Subjects:
Online Access:https://animorepository.dlsu.edu.ph/etd_masteral/3001
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: De La Salle University
Language: English
Description
Summary:Tagalog is a free-word order language, thus, exhibiting the problem of discontinuous constituents. Current approaches to resolve this type of problem use the configurational (or hierarchical) syntactic structure as a model to represent surface ordering of constituents, which is suitable for fixed-word order languages. The presence of discontinuous constituents in sentences represented in hierarchical structures also causes parsing delays due to constituent movements and backtracking. TagFWO (Tagalog Free-Word Order parser) is a web-based implementation of a new technique to address the problem of discontinuous constituents. It flattens the hierarchical syntactic structure (that is, the verb and all the rest of the constituents are sisters) thus, processing can be done in a single step without loss of semantic information. It uses the Head-driven Phrase Structure Grammar (HPSG) formalism to encode the linguistic knowledge of Tagalog. The TagFWO parser was tested and evaluated using Tagalog free-word order sentences (grammatical and ungrammatical). The results showed that the grammatical (which should be accepted) sentences are successfully parsed, and ungrammatical sentences are rejected and outputted an error message. The results show that the new technique is appropriate for Tagalog, and requires less computing time in contrast to other existing approaches.