个人教育数据的敏感性识别与隐私计量研究

Research on Sensitivity Identification and Privacy Measurement of Personal Education Data

[目的/意义]现行教育数据行业标准将个人教育数据划分为三个级别,但缺乏分级定量研究支撑,文章定量测度教育隐私值,解决该问题。[方法/过程]首先,依据隐私主体敏感认知,归纳文本类型,建立教育隐私文本库,作为敏感数据识别来源;其次,识别教育敏感数据单元,设计敏感词表建立流程,建立教育敏感词表;最后,分析隐私影响因素,形成计量指标,构建计量模型,计量教育隐私。[结果/结论]敏感性识别结果为个人教育敏感数据项,包括个人基础数据(标识数据、半标识数据、鉴别数据、联系数据、健康数据)、奖惩数据(奖励数据、惩罚数据)、学生管理数据(学籍数据、读者数据、资助数据、就业数据)、学习数据(课程数据、综测数据)4类。隐私计量结果表明,个人基础数据的隐私值最高,奖惩数据的隐私值次之,学生管理数据和学习数据的隐私值并列最低。科学权威的教育数据等级体系分类表缺失和隐私文本的切题性不高,致使教育敏感词表和隐私文本库的质量欠佳,可能影响计量结果的准确度,但尽管如此,样本统计结果还是比较全面反映了该领域的真实情况,体现了教育数据项的重要程度。

[Purpose/significance]The current industry standards for education data divided personal education data into three levels,this paper quantitatively measures education privacy values to solve the problem that this classification standard lacks quantitative research support.[Method/process]Firstly,based on the sensitive perceptions of privacy subjects,we summarize the types of privacy texts and establish an education privacy text database as the identification sources of sensitive data.Secondly,we identify education sensitive data units and design a process of sensitive vocabulary to build an education sensitive vocabulary.Finally,we analyze the influencing factors and form measurement indicators,constructing a privacy measurement model to measure education privacy.[Result/conclusion]The sensitivity identification results are personal education sensitive data items,including personal basic data(identification data,semi-identification data,identity authentication data,contact data,health data),reward and punishment data(reward data,punishment data),student management data(student status data,reader data,financial aid data,employment data),and learning data(course data,comprehensive test data).According to the ranking results of privacy values,personal basic data is the highest,reward and punishment data is the second,student management data and learning data are tied lowest.A scientific and authoritative classification system for education data is lacking and the privacy texts have low relevance,resulting in poor quality of education sensitive vocabulary and privacy text database,which may affect the accuracy of measurement results,but despite this,the sample statistical results still comprehensively reflect the real situation in the field and the importance of education data items.