Year: 2019
Type: dissertation
Abstract: Chapter 1 introduces a Thai word segmentation problem and its role in an analysis of a brand crisis.Word boundary ambiguity has been a challenge in Thai language processing.Incorrect word segmentation may result in misleading interpretations.Chapter 2 explains Thai language fundamentals that are related to word segmentation.The chapter describes word formation, which is a sequential combination of words that form a new word.The roots of a compound word may have different meanings or can be interpreted differently from the word.Because of this difference, word segmentation may not produce a meaning that is similar to the meaning of the whole word, making the outcome ambiguous.Chapter 3 proposes word segmentation rule and two post-processing algorithms to the existing machine-learning model, a Conditional Random Fields (CRF).The two proposed algorithms are word-merging and word-splitting algorithms.CRF is one of the most accurate word segmentation models among Thai word segmentation methods.The existing CRF-based word segmentation model was trained on Benchmark for Enhancing the Standard of Thai Language Processing (BEST2009) corpus developed by National Electronics and Computer Technology Center.The first problem is that the corpus does I would like to express my deepest gratitude to my chief advisor, Dr. Hitoshi Okada, for his excellent guidance, caring, patience, and providing me with an excellent atmosphere for doing research.His patience and support helped me overcome many crisis situations and finish this thesis.
Authors: NARARATWONG, Rungsiman, ナララウォン, ランシマン, NARARATWONG , RUNGSIMAN
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7006267755