基于神经网络技术的评卷误差控制模型及其应用  

Marking error control model Based on neural network technology and its application

作　　者：丁文[1] 杨卫东[1] 刘继来[1] 

机构地区：[1]浙江省高校招生办公室,浙江杭州310012

出　　处：《浙江工业大学学报》2003年第4期419-423,431,共6页Journal of Zhejiang University of Technology

摘　　要：各类考试的评卷工作中 ,如何使评卷人员在整个评卷过程中始终掌握评分标准的一致性和评卷的稳定性 ,控制评卷误差 ,是有关方面一直在关注和急待解决的问题。将评卷的模式从传统的纸上评卷改为在计算机上进行的网上评卷 ,把智能神经网络技术应用到网上评卷中 ,用柯尔莫哥洛夫—斯米尔诺夫检验作为控制 ,通过检验输出的值是否符合要求 ,达到随时掌握和及时调整评卷人员的状态 。How to maintain consistence and stability and prevent errors in the process of marking is a problem which has been attracting attention from people concerned and which requires immediate solution in the marking of examination papers. To guarantee the marking quality, a method is suggested as follows: to change traditional paper marking into net marking on the computer, to apply intelligence neural network technology to the net marking,and to monitor by employing Kolmogorov\|Smirnov test, to test whether the output result meets the criteria so that the mentality of examination paper markers can be learned and adjusted promptly.

关 键 词：神经网络 评卷误差控制模型 应用 柯尔莫哥洛夫-斯米尔诺夫检验 评卷质量 HOPFIELD模型 网络教学 

分 类 号：G424.74[文化科学—课程与教学论] TP183[文化科学—教育学]