基于聚类分析技术的教师识别过程研究  

Teacher recognition process research with clustering analysis technology

作　　者：陈春燕[1] 叶枫[1] 

机构地区：[1]蚌埠医学院卫生管理系,安徽蚌埠233000

出　　处：《淮南师范学院学报》2017年第2期122-124,共3页Journal of Huainan Normal University

基　　金：安徽省高校人文社科研究重点项目(SK2017A0182)

摘　　要：分析合肥10所学校学生的考试试卷,从试卷知识点和试卷成绩两个方面共提取5类特征,使用这5类特征进行教师识别,进而验证教师对学生成绩的影响。从试卷成绩方面提取3类特征,分别是班级均得分率*试题难度、得分率分段比例、错选项比例;从试卷知识点方面提取2类特征,分别是班级均掌握情况、知识点/抽象能力掌握情况分段比例。将这5类特征作为特征向量,使用聚类分析的方法进行教师识别,实验结果表明识别正确率达到73%。The paper analyzed the students＇test papers in ten schools in Hefei. Five types of features were extracted from testing scores and examination paper knowledge. The five types of feature were used to proceed teacher recognition, judge whether teachers had an impact on students＇achievement. The three types of features were mean scoring rate of class＊item difficulty, scoring rate segment ratio and wrong option ratio from testing scores. The two types of features were class mean mastery and knowledge point mastery segment ratio from examination paper knowledge. Taking the five types of features as feature vector, the author used clustering analysis technology to proceed teacher recognition. The experimental results showed that the recognition accuracy reached 73%.

关 键 词：聚类分析 教师识别 特征提取 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]