CORPUS ANALYSIS WITH THE AID OF MACHINE READABLE DICTIONARY-in the Case of Texts Written in Indic Scripts-

这个项目的目的是开发这样一种方法，通过这种方法，计算机可以检索语言语料库，并在适当的机器可读词典(MRD)的帮助下产生有语法标记的文本。该项目的目标语言是印地语、梵语、泰语、柬埔寨语和缅甸语，所有这些语言都有以下共同特点。采用印度语脚本，可以制定使用ASCII码的通用转录规则。被归类为所谓的SOV类型语言，在这种语言中，出现在运行文本中的任何单词形式都可以分为基本和后缀，即使是粘合的或屈折的。我们为目标语言设计了通用的MRD格式，并开发了几个在MS-DOS和Unix OS上处理MRD的程序。每个MRD包含超过10,000个中心词。至于印地语，由于该项目的完成，以下网站提供了一些暂定服务。日语版(http：//www3.aa.tufs.ac.Jp/-kmach/index.htm)

The aim of this project is to develop such a method by which a computer could retrieve language corpus and produce grammatically tagged texts with the aid of proper Machine Readable Dictionaries (MRD). The target languages of the project are Hindi, Sanskrit, Thai, Cambodian and Burmese, All these languages have the following features in common.adopting Indic scripts, for which common transcription rules using ASCII code can be made up.being classified to so-called SOV type languages, in which any word form occurring in running texts can be divided into BASE and SUFFIX(es), even if agglutinative or inflectional. We have designed the common format of MRD for the target languages and developed severalprograms which work with MRDs both on MS-DOS and Unix OS.Each MRD contains more than 10,000 head words. As for Hindi some tentative services due to the achievement of this project are available on the following web site.Japanese version (http : //www3.aa.tufs.ac. jp/-kmach/index.htm)