Noun Phrase Engine for Thai NLP
Nuchjarin Krukaset, Watcharapong Krukaset, Chouvalit Khancome · 2024
The core objective of Natural Language Processing (NLP) is to equip computers with the capability to comprehend and analyze human communication. In the context of the Thai language, characterized by its structured clarity, extensive research has been ongoing in this domain. This study introduces NPE (Noun Phrase Engine), a novel software engine tailored for managing Thai noun phrases. Operating on a directional graph framework governed by 17 main patterns and accommodating up to 147 sub-level patterns, NPE integrates algorithms for word segmentation, noun phrase analysis, and noun phrase generation. Experimental results demonstrate NPE's exceptional performance, achieving 100% accuracy in word segmentation compared to established tools like Longest, Multi-cut, and Newmm. Furthermore, NPE exhibits comparable or superior performance in noun phrase analysis when evaluated against Generative AI models such as ChatGPT version 3.5, Gemini, and Alisa, achieving 100%accuracy. Additionally, in noun phrase generation, NPE outperforms Unigram, Bigram, and Trigram algorithms in the PyThaiNLP library, producing more accurate noun phrases significantly.