摘要
为有效判别高校学生在线作业中复制、代写、抄袭等不良现象,引入击键行为识别技术。根据不同应用场景构造了单字符间特征值、关键字字符串内特征值以及关键字字符对特征值,并结合神经网络算法构建了一种基于MC-LSTM的在线作业击键行为识别模型。以IT课程“数据库原理与应用”为例,设计并实现了基于MC-LSTM的在线作业行为识别系统,并测试了不同特征值及特征值组对识别效果的影响。实验结果表明,该系统能够有效判别在线作业自主完成情况,从而对在线教学质量提供保障。
In order to effectively identify undesirable phenomena such as copying, ghostwriting, and plagiarism in college students’ online homework, the keystroke behavior recognition technology is introduced. According to different application scenarios, the eigenvalues between single characters, within keyword string and keyword character pair are constructed. Combined with the neural network algorithm, an online homework keystroke behavior recognition model based on MC-LSTM is established. Taking the IT course “Database principles and applications” as an example, an online homework autonomous discrimination system based on MC-LSTM is designed and implemented, and the effects of different eigenvalues and eigenvalue groups on the recognition effect are tested. The experimental results show that the system can effectively distinguish the autonomous completion of online homework, so as to provide guarantee for the quality of online teaching.
作者
祝锡永
吴炀
ZHU Xiyong;WU Yang(School of Economics and Management,Zhejiang Sci-Tech University,Hangzhou 310018,China)
出处
《实验技术与管理》
CAS
北大核心
2022年第2期198-204,共7页
Experimental Technology and Management
基金
国家自然科学基金项目(71501172)
浙江省自然科学基金项目(LY18G020017)。
关键词
神经网络
在线作业
自主判别系统
行为识别
neural network
online work
autonomous discrimination system
action recognition