论文部分内容阅读
在简述词形归并(原形化处理)基本目标的基础上,着重分析UMLS专家词典的构建方式与词典工具的核心功能,以及Norm原形化工具的处理机制;在此基础上,提出一种词形归并算法及Norm处理缺陷的修正办法,并收集医学词表数据进行算法测试与验证;此外,本文算法与经典的Porter算法进行了多方面的深入比较研究。
On the basis of briefly describing the basic goals of morphological merging (prototyping), this paper focuses on analyzing the construction of lexicon of UMLS experts and the core functions of lexical tools, and the processing mechanism of Norm prototyping tools. On this basis, Algorithm and Norm processing defect correction method, and collect medical vocabulary data for algorithm testing and verification; In addition, the algorithm and the classic Porter algorithm conducted in-depth comparative study in many aspects.