Mahidol University Logo
Faculty of ICT, Mahidol University
 

Admissions

Printable Version

 

THAI SPELLING CORRECTION USING STATISTICAL LANGUAGE MODEL

 

TITLE THAI SPELLING CORRECTION USING STATISTICAL LANGUAGE MODEL
AUTHOR PORNCHAI TUMMARATTANANONT
DEGREE MASTER OF SCIENCE PROGRAMME IN COMPUTER SCIENCE
FACULTY FACULTY OF SCIENCE
ADVISOR DAMRAS WONGSAWANG
CO-ADVISOR THANWADE SUNETNANTA
 
ABSTRACT
The research topic of Thai spelling correction is an important contribution to Thai text recognition research. These two research areas both share the same objective of achieving high recognition accuracy. Thai spell checking is normally achieved by comparison with dictionaries. However, language grammars and other related rules are not able to completely deal with certain specific language usages. This research work proposes a Thai spelling correction prototype using the Statistical Language Model, a technique that estimates the distribution of various natural language phenomena which is considered to be characteristics of Thai language. The technique of proposed prototype is able to check spelling of words based on the meaning within a sentence as well as to recognise non-standard written language behaviour. The spelling correction works on both word and character levels The proposed prototype has been tested using from the Document Image Corpus, containing outputs of the Thai Optical Character Recognition program. The test data includes 1,442 text files of 19.5 megabytes. Experimental results show that the proposed technique can improve recognition rates by 18% using documents from Optical Character Recognition with recognition rates of 75% or more.
KEYWORD SPELLING CORRECTION / POST PROCESSING / STATISTICAL LANGUAGE MODEL

 

Go to Top

 

ICT Building, Mahidol University, 999 Phuttamonthon 4 Road, Salaya, Nakhonpathom 73170 Tel. +66 02 441-0909 Fax. +66 02 849-6099
Mahidol University Computing Center, The Faculty of ICT, Mahidol University , Rama 6 Road, Rajathevi, Bangkok 10400 Tel. +66 02 354-4333 Fax. +66 02 354-7333