CAJ | 학술논문

名实体歧义是机器对自然语言进行理解时经常遇到的问题，为使机器能够正确地分析自然语言文本，对名实体消除歧义亟待解决。近年来，随着Wikipedia等语义知识库的出现，大量基于知识库的消歧方法被提出。命名实体消歧的任务是将文本中具有多个含义的实体指称去除歧义，并将其链接到知识库中的唯一实体。本文采用 DBpedia作为知识库，基于图的方法进行实体消歧。
명실체기의시궤기대자연어언진행리해시경상우도적문제，위사궤기능구정학지분석자연어언문본，대명실체소제기의극대해결。근년래，수착Wikipedia등어의지식고적출현，대량기우지식고적소기방법피제출。명명실체소기적임무시장문본중구유다개함의적실체지칭거제기의，병장기련접도지식고중적유일실체。본문채용 DBpedia작위지식고，기우도적방법진행실체소기。
Ambiguity is one of the most common problems in natural language processing.In order to make machine analy-sis natural language texts correctly, eliminating ambiguity is an urgent problem to be addressed.In recent years, with the e-mergency of knowledge base such as Wikipedia, there are large amount of method proposed based on knowledge base.The task of named entity disambiguation is to eliminate ambiguity for the mentions which has multiple meanings, and link it to only one entity in knowledge base.This article uses a graph based method, and employs DBpedia as the knowledge base to link.