Syllable Segmentation Algorithm for Myanmar Language

Cho Cho Hnin, Naw Naw

Abstract: Myanmar language does not have word boundary or white space between words. Thus, it is problematic to tokenize these words into the meaningful words before the process of text mining. There are many word segmentation methods using Unicode standard encoding for Myanmar language. Many people still use Myanmar Zawgyi-One font especially in social media contents. Thus, it is relevant to focus on the word segmentation for different social media text mining. Some Myanmar informal texts, especially in social media contents, can contain English words among Myanmar words. In such case, it is necessary to segment both of Myanmar words and English words from these mixed informal texts. This paper proposes a syllable segmentation algorithm for Myanmar text (Zawgyi-One Standard) and for the text with the combination of Myanmar words and English words.

Keywords: Natural language processing, Myanmar language, word segmentation, syllable segmentation