Head of Sub-Project 2

Associate Professor Wirote Aroonmanakun, Ph.D.

Principles and Reasons

Compounding is an important process in creating a new word in Thai. Recognizing a compound is an active research in natural language processing. A lot of errors in Thai word segmentation are caused by unidentified compounds in the text. A structural analysis of Thai compounds in a corpus will reveal some basic facts useful for understanding this problem.

Previous research on term extraction and compound extraction in various languages will be reviewed. Various methods including rule-based and statistical-based used to handle these issues will be summarized. A method that is suitable for Thai compound extraction will be proposed, implemented and evaluated.

This project is part of the Thai language processing project, in which the module created from this project along with other modules like word segmentation, named entity recognition and part-of-speech tagging will be included in a program for analyzing Thai words.

Expected Benefits

It will contribute to the creation of a Thai word analysis program, which is similar to the "WordSketch" program used for analyzing all information relevant to a word. This program will be useful not only for Thai language processing, but also for Thai lexicography and Thai language teaching.