中文摘要：

藏语词类的基本属性描述是藏语语料库多级加工的基础，根据藏语语料库多级加工的实际需要，通过对藏语词类的语法、语义信息的描述和藏语词语后添加方式的分析，可以为计算机对藏文的词语搭配结构、语法信息、语义信息和词类标记进行自动处理提供一个切实可行的方法。

英文摘要：

Tibetan corpus multistage process is based on description of the basic attributes of the Tibetan lexicon,according to the actual need of the Tibetan corpus multistage process,through the syntax and semantics of Tibetan information description and Tibetan words after the analysis,it can provide a feasible method for the computer to process automatically the Tibetan collocations structure,grammatical information,semantic information and POS-tagging.