Improving Thai word segmentation with Named Entity Recognition

Tepdang, Sayan; Haruechaiyasak, Choochart; Kongkachandra, Rachada

doi:10.1109/iscit.2010.5665124

Cited by 7 publications

(1 citation statement)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In Thai, there are many proposed techniques for extracting information and relations such as Aunanan et al (2016) proposed extracting Thai news contents by auto-tagging for faster information retrieval, and Sutheebanjard and Sutheebanjard (2009) proposed the method to extract Thai personal named entities without using Word Segmentation or POS Tagging from news articles which greatly reduces the time and effort used in building the training corpus, while, Tongtep and Theeramunkong (2009) presented a feature-based approach for extracting relations between a pair of named entities (person name, location, and action) from Thai news documents, which gave accurate results. The major problems of information extraction from Thai text is caused by long sentences with no use of punctuation and the lack of use of upper and lowercase letters to identify named entities, and the inconsistency of the writing (Tepdang et al, 2010;Wikaha and Netisopakul, 2014).…”

Section: Related Workmentioning

confidence: 99%

Historical Relation Extraction from Buddhist Temple Documents of the Lanna Kingdom

Chantaraj

Rungrattanaubol

Na-udom

2019

Journal of Computer Science

View full text Add to dashboard Cite

A substantial amount of valuable knowledge has been recorded, namely historical documents, in the form of unstructured data. The Extraction of such information will provide useful historical knowledge. The objective of this study, is to propose a relation extraction technique of unstructured data from Buddhist temple documents in the Lanna Kingdom, consisting of relationships including the name of the kings who built the temple, the name of the king who created or restored a historical site within its grounds and the Buddhist era the temple was built. The research process began with collecting temple data located in the upper northern part of Thailand, with published documentation of temple history in electronic form from Temples Thailand online website. The data was transformed into an electronic document or a text file, then passed to the LEXTO program for word segmentation, while retaining the original meaning. We designed and developed an algorithm to extract relationships between named entities from Buddhist temple documents, and the application of information extraction principles to the process. We then evaluated the 1,093 original documents by using the algorithm to extract the required information. From the research, the results indicated that the proposed technique can be applied effectively, to accurately extract the relations related to Lanna kingdom history from Buddhist temple documents.

show abstract

Section: Related Workmentioning

confidence: 99%