Building a Spelling Checker for Documents in Khmer Language

The practice of Khmer language teaching at Tra Vinh University shows that students often deal with many difficulties when making reports on subjects of Khmer language. The need to print, edit Khmer text with MS Word by using conventional typing on the ASCII keyboard and the Unicode fonts available made students have many spelling errors, ranging from syllable level to vocabulary level, the syntax of the sentence. Students have not correctly identified the various syllables in word structures, encountered errors in pronunciation and spellings in Khmer. In fact, Khmer-Vietnamese community in Vietnam has few people, scattered, culturally and economically limited. Moreover, Khmer writing system is complex, from the rich alphabet to the multi-syllable structure of the syllable, with no whitespace separated from the sentences. There are many borrowed words which have changed much from the spoken words in Cambodian origins, so the users face many difficulties. These reasons have been a major obstacle in the process of dealing with Khmer language in Vietnam. So far, there have been no solutions to meet the demand for spelling checking in the Khmer scripts. Based on the actual analysis of the students' common misspellings, it can be proposed a solution to the application of the syllable model, the syllabus processing algorithm using the Khmer vocabulary that has been authored by the authors. Previously built KSC (KhmerSpellChecker), an automatic spell checker, integrated into MS Word, helps learners to compose Khmer texts. The approach is feasible. KSC was installed and test results were tested by Khmer language experts at Tra Vinh University for high accuracy.