基于文本相似性检索技术解决命题中重题检测问题的实践——以北京市自学考试命题为例  

Implementation of Duplicate Items Checking Based on Text Similarity Detection:A Case Study in Self-taught Examination in Beijing

作　　者：沈钢[1] 

机构地区：[1]北京教育考试院,北京100083

出　　处：《中国考试》2018年第3期38-42,共5页journal of China Examinations

摘　　要：本文阐述试题查重系统的设计原理、基本框架和工作流程,以北京市自学考试命题数据为基础,进行试题查重系统测试并分析数据结果。实验结果表明,试题查重系统在重题和相似题的判断上具有很高的准确率和查全率,可以有效地控制自学考试复本试卷间的试题重复率。This article starts with the puzzles about duplicate items checking, with the concept of text similarity detection introduced, illustrates the design principle, frame and workflow of duplicate items checking system. The system is tested with the data of Self-taught Examination in Beijing. The result shows the system has a high level of precision and recall in duplicate items checking, which means it can control the repetition rate between parallel forms in Self-taught Examination effectively.

关 键 词：重题检测 文本相似性检索 自学考试 命题 

分 类 号：G405[文化科学—教育学原理]