[1]冷强奎,刘雨晴,秦玉平.基于二元模糊匹配的编程题智能评分方法[J].计算机技术与发展,2020,30(02):71-74.[doi:10. 3969 / j. issn. 1673-629X. 2020. 02. 015]
　LENG Qiang-kui,LIU Yu-qing,QIN Yu-ping.Intelligent Scoring Method for Programming Problems Based on Binary Fuzzy Matching[J].COMPUTER TECHNOLOGY AND DEVELOPMENT,2020,30(02):71-74.[doi:10. 3969 / j. issn. 1673-629X. 2020. 02. 015]
点击复制

基于二元模糊匹配的编程题智能评分方法()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
30
期数:
2020年02期
页码:
71-74
栏目:
智能、算法、系统工程
出版日期:
2020-02-10

文章信息/Info

Title:
Intelligent Scoring Method for Programming Problems Based on Binary Fuzzy Matching
文章编号:
1673-629X(2020)02-0071-04
作者:
冷强奎1; 刘雨晴1; 秦玉平2
1. 渤海大学 信息科学与技术学院,辽宁 锦州 121000; 2. 渤海大学 工学院,辽宁 锦州 121000
Author(s):
LENG Qiang-kui1; LIU Yu-qing1; QIN Yu-ping2
1.School of Information Science and Technology,Bohai University,Jinzhou 121000,China; 2.School of Engineering,Bohai University,Jinzhou 121000,China
关键词:
二元模糊匹配; 自动评分; 词频统计; 余弦相似度; 向量空间模型
Keywords:
binary fuzzy matching; automatic scoring; word frequency statistics; cosine similarity; vector space model
分类号:
TP391
DOI:
10. 3969 / j. issn. 1673-629X. 2020. 02. 015
摘要:
针对传统编程题自动评分方法不能准确衡量学生程序与参考答案之间的接近程度等问题,提出了一种基于二元 模糊匹配的编程题智能评分方法。 第一元为结构匹配,检测学生程序在变量声明、数据输入、函数调用、控制结构等方面 与参考答案的相似程度,目的是快速判断学生程序中是否存在关键的采分点。 第二元为词语匹配,首先进行词频统计,以 确定每个词的权重。 然后,分别构造学生程序与参考答案的向量空间模型,并计算两者的余弦相似度,来作为评判词语相 似的依据。 最终的分数由结构相似度和词语相似度的加权分数计算得出。 由于该二元匹配方法不关心程序中结构/词语 的先后顺序和是否为精确表达,因此被称为是模糊的。 仿真实验表明,该方法具有不错的评分准确性,尽管与人工评分相 比还存在一定的差距,但当试题规模较大时,可以作为人工评分的替代手段。
Abstract:
Aiming at the problem that the traditional automatic scoring method for students’ programming assignments can not accurately measure the proximity between the student program and the answer,a method of intelligent scoring based on binary fuzzy matching (BFM) is proposed. The first operation of BFM is called structural match,which detects the similarity of the student program to the answer in terms of variable declaration,data input,function call,control structure,etc.,and aims to quickly determine whether there is a key score point in the student program. The second operation of BFM is called word match. The word frequency statistics are first performed to determine theweight ofeach word. Then,thevectorspacemodelsofthestudent program and theanswerareconstructed respectively. Next,the cosine similarity is calculated as the basis for judging the similarity of the words. The final score is synthesized by the separate weighted scores of structural similarity and word similarity. Since the binary matching method does not care about the order of structures/words in the program and whether these terms are accurately expressed,it is said to be fuzzy. Simulation experiments show that the method has excellent scoring accuracy. Although there is a certain gap from the manual scoring,when the scale of programs is large,it can be used as an alternative to manual scoring.

相似文献/References:

[1]付鹏斌,刘 曼,杨惠荣.结合学科情感分析与依存关系的相似度评分[J].计算机技术与发展,2022,32(02):32.[doi:10. 3969 / j. issn. 1673-629X. 2022. 02. 005]
　FU Peng-bin,LIU Man,YANG Hui-rong.Similarity Score Combining Subject Sentiment Analysis and Dependency Relationship[J].COMPUTER TECHNOLOGY AND DEVELOPMENT,2022,32(02):32.[doi:10. 3969 / j. issn. 1673-629X. 2022. 02. 005]

更新日期/Last Update: 2020-02-10