语句相似度计算在主观题评判中的应用  

Application of sentence similarity measurement in automated assessment of subjective tests

作　　者：于淼[1] 王日宏[1] 

机构地区：[1]青岛理工大学计算机工程学院,山东青岛266033

出　　处：《计算机应用》2008年第12期3254-3256,3260,共4页journal of Computer Applications

基　　金：建设部研究开发项目(06-k9-15)

摘　　要：在网络考试系统中对主观题自动评判运用自然语言识别技术是计算机领域的一个研究热点。结合中文分词方法,在基于《知网》知识库的词语相似度算法的基础上,运用数学上求解最优指派问题的匈牙利算法建立句子和句群的相似度计算模型,并根据句子相似度值给出主观题正确分数评判。Trying to use the natural language in the automated assessment of network examination system has become a research hotpot in computer area. Referring to Chinese word segmentation, the sentence similarity calculation model was established by using the model of words semantic similarity based on HowNet and sentence semantic similarity model based on Hungary algorithm. In accordance with the sentence similarity, the correct value was evaluation scores for automated assessment.

关 键 词：自动评判 中文分词 相似度模型 匈牙利算法 

分 类 号：TP311.52[自动化与计算机技术—计算机软件与理论]