CAJ | 학술논문

采用统计方法来识别中文四字姓名。该方法将中文四字姓名的识别过程分为姓名候选和姓名求精两个阶段。采用二元隐马尔科夫模型从已经切分好的文本中候选姓名。利用边界规则对候选姓名进行求精。实验结果表明，该方法的召回率为82．9％，准确率为87．3％。
채용통계방법래식별중문사자성명。해방법장중문사자성명적식별과정분위성명후선화성명구정량개계단。채용이원은마이과부모형종이경절분호적문본중후선성명。이용변계규칙대후선성명진행구정。실험결과표명，해방법적소회솔위82．9％，준학솔위87．3％。
Automatic recognition of Chinese four character name is an important part of Chinese Named Entity recognition. The Chinese fourcharacter name recognition process can be divided into two stages, namely, the name candidate and the name of refinement. A Hidden Markov Model （HMM） is applied for the extraction of candidate names from text. The boundary rule on candidate names is used for the refinement. The test experiments show that the precision and recall rate reach 87.3% and 82.9% respectively.