Panel: Good Spelling of Vietnamese Texts, One Aspect of Computational Linguistics in Vietnam

Good Spelling of Vietnamese Texts, one aspect of computational linguistics in Vietnam PHAN Huy Khanh Department of Information Technology DaNang University 17, Le Duan Street, DaNang City, Vietnam phanhuykhanh@dng.vnn.vn Abstract There are many challenging problems for Vietnamese language processing. It will be a long time before these challenges are met. Even some apparently simple problems such as spelling correction are quite difficult and have not been approached systematically yet. In this paper, we will discuss one aspect of this type of work: designing the so-called Vietools to detect and correct spelling of Vietnamese texts by using a spelling database based on TELEX code. Vietools is also extended to serve many purposes in Vietnamese language processing. Introduction For the past two decades computational linguistics (CL) has progressed substantially in Vietnam, mainly in these basic aspects: data acquisition from the keyboard, encoding, and restitution through an output device for Vietnamese diacritic characters, updates on the fonts in Microsoft DOS/Windows, standardization for Vietnamese (James Do, Ngo Thanh Nhan), automatic translation of English documents into Vietnamese and vice versa (Phan Thi Tuoi, Dinh Dien), recognition handwriting (Hoang Kiem, Nguyen Vanof Khuong), speech processing (Nguyen Thanh Phuc, Quach Tuan Ngoc), building bilingual dictionaries such as English-Vietnamese and V- E, French-Vietnamese and V-F dictionaries (Lac Viet), archives of old Sino-Vietnamese documents (Ngo Trung Viet, Cong Tam), etc. Some of these works have been presented in Informatics and IT workshops organized in Vietnam. These efforts are modest and do not yet show our full potential. There are many reasons for this weakness. The major reasons that the different efforts are quite isolated and there is not enough coordination. Some coordinated workshops held from time to time would be very helpful. At the IT Dept. DaNang University we are building a lexical database based on TELEX code for accomplishing the following tasks: - Converting Vietnamese texts from any font to any other font. - Putting texts in alphabetical order independently of the font in use. - Looking up words up in the monolingual and / or multilingual dictionary. - Building specialized monolingual dictionaries. At present, we are taking part in the GETA, CLIPS, IMAG, France, in the FEV project: for a multilingual dictionary: French-Vietnamese via English. In fact, inputting Vietnamese texts still encounters many problems, not yet solved properly. The most common mistakes in detecting and correcting spelling errors are: - wrong intonation or misspelling, - not following spelling specialization, not using syllables systematically in the same texts, etc. Winword, a commercial text processor, is not able to detect and correct spelling mistakes. The program designed by Ngo Thanh Nhan (without an associated spelling dictionary) and other software packages for Vietnamese still do not offer adequate solutions. We propose here a general solution for building the so-called Vietools for detecting and correcting spelling errors. Vietools is designed for office application such as Winword, Excel, Acess, PowerPoint, etc. in Microsoft Windows. Vietools has also been extended for converting and rearranging Vietnamese words in the dictionaries and consulting the Vietnamese …

Content truncated to 500 words