代漢語語義詞典多義詞詞庫的校正和再修訂

本文依據面向漢語信息處理的詞語義區分的完備性和操作性原則，基于代漢語語義詞（SKCC和代漢語語法信息詞（GKB）的詞對，以語料依托並結合代漢語詞、義詞詞林和代漢語搭配詞詞等詞資源進行代漢語語義詞的多義詞詞校工作。首先SKCC和GKB的詞對入手，設了候選擬修改多義詞的取算法，對取出來的1605個多義詞進行了義補錄合併刪除，補充釋義，修改翻譯和示例等工作，時針對類特殊的食物+作物類多義詞建立了義的樹形結構以滿足粒度的詞義消歧任要求。方便續的修改工作，本文開發了SKCC和GKB之間的詞映射，在的基礎進行了多義詞映射工作。

This paper is rooted in the two principles and methods that should be followed by sense discrimination for Chinese language processing: Completeness and discreteness. Built on the comparison of Semantic Knowledge-base of Contemporary Chinese (SKCC) and Grammatical Knowledge base of Contemporary Chinese (GKB), supported by large scale corpus, we conducted our new editing and checking works. Firstly, we designed a novel multi-sense lexicon candidate abstraction algorithm based on lexicon comparison between SKCC and GKB. For all 1605 candidate multi-sense lexicon, we conducted editing work on the senses, explanation, and its translation. Then, we built a tree structure to process a special food and plant lexicon. Thirdly, a mapping platform between SKCC and GKB has been built to help us built mapping relationships between multi-sense lexical between SKCC and GKB. Finally, we finished mapping work for all multi-sense lexicon in SKCC.