A Unified Model of Thai Romanization and Word Segmentation

A Unified Model of Thai Romanization and Word Segmentation Wirote AROONMANAKUN Dept. of Linguistics Chulalongkorn University Bangkok 10330, Thailand Wirote.A@chula.ac.th Wanchai RIVEPIBOON Dept. of Computer Engineering Chulalongkorn University Bangkok 10330, Thailand Wanchai.R@chula.ac.th Abstract Thai romanization is the way to write Thai language using roman alphabets. It could be performed on the basis of orthographic form (transliteration) or pronunciation (transcription) or both. As a result, many systems of romanization are in use. The Royal Institute has established the standard by proposing the principle of romanization on the basis of transcription. To ensure the standard, a fully automatic Thai romanization system should be publicly made available. In this paper, we discuss the problems of Thai Romanization. We argue that automatic Thai romanization is difficult because the ambiguities of pronunciation are caused not only by the ambiguities of syllable segmentation, but also by the ambiguities of word segmentation. A model of automatic romanization then is designed and implemented on this ground. The problem of romanization and word segmentation are handled simultaneously. A syllable-segmented corpus and a corpus of word-pronunciation are used for training the system. The accuracy of the system is 94.44% for unseen names and 99.58% for general texts. When the training corpus includes some proper names, the accuracy of romanizing unseen names was increased from 94.44% to 97%. Our system performs well because it is designed to better suit the problem. 1 Introduction The attempt to create a system of Romanization for Thai texts began since the 17th century by French missionaries (Griswold, 1960). However, the work was left unnoticed by other foreigners, who tend to romanize Thai words using their own languages notations. Issues of standardizing Thai romanization have been concerned again in the early 20th century (Frankfurt (1906), Petithuguenin (1912),Vajiravudh (1912,1931), Frankfurt et al. (1931)). Some systems of romanization are done on the basis of orthographic form, e.g. the system proposed by the ISO (ISO 11940 : 1998). Some systems are based on the pronunciation, such as the Royal Institute's system, Thiengburanathum's system (in his Thai-English dictionaries). Some are not totally based on orthographic form or pronunciation, e.g. the system proposed by King Rama VI.1 This could explain why there are many ways to romanize a Thai word. For example, a common name like “เกรียงศักดิ์” can be romanized as “Kriangsak”, “Kriengsakdi”, “Kriengsak”, or “Kreangsak”. To lessen this problem, the Royal Institute has established the principle of romanization for Thai.2 The principle has been endorsed by the United Nation (UN 2002) and slightly modified by the American Library Association and Library of Congress (ALA-LC 1997) for romanizing Thai scripts. Though the standard has been established, it is still not easy for general users to do romanization by hand. People tend to romanize Thai words on their own, rather than adhering to the principle. To help promoting the principle, Thatsanee et al. (1999) proposed an idea to develop an automatic romanization system. We agree with Thatsanee et al. that an automatic Thai romanization system is necessary for ensuring the standard. Such …

Content truncated to 500 words