一种基于知网的主观题阅卷算法  

A Grading Algorithm for Subjective Tests Based on HowNet

作　　者：丁振国[1] 陈海霞[1] 

机构地区：[1]西安电子科技大学计算机学院,陕西西安710071

出　　处：《微电子学与计算机》2008年第5期108-109,113,共3页Microelectronics & Computer

基　　金：国家"八六三"计划项目(2004AA1Z2520)

摘　　要：通过分析知网中词语表示方法提出了一种新型的义原相似度计算方法进而获得词语相似度的计算公式,以词语相似度计算为基础提出了主观题阅卷算法的设计与实现.实验结果证明,该算法能够较好的评阅学生答案中语义正确的内容,这在提高系统阅卷的正确性方面具有一定的意义.Through analyzing the expression of words in HowNet, a new method to calculate the similarity betweens sememes was proposed, then the calculating way of similarity between words was got in this paper. Based on the words similarity calculating the paper designs and implements a Grading algorithm for subjective tests. The exPerimental result shows that the algorithm can better grade the semantic right content in student＇s answers which has significance in improving the correctness of grading paper.

关 键 词：知网 主观题 阅卷算法 相似度 

分 类 号：G434[文化科学—教育学]