摘要
[目的/意义]现行金融数据行业标准规范中,依据“遭到破坏或泄露产生的影响程度”将个人金融数据隐私分为3个或4个级别,但均缺乏定量研究支撑。挖掘金融隐私相关文本,定量计算金融数据项的隐私值,为个人金融隐私分级提供科学客观依据。[方法/过程]普查金融隐私相关文献,建立金融隐私文本库;依据金融数据行业标准规范,建立金融隐私词表框架,采集金融敏感词汇;调查金融机构,采集客户个人数据,完善金融敏感词汇;基于文本句法结构,从文本库中析出金融隐私动词和金融敏感程度词,建立金融隐私词表;基于隐私词表,抽取金融文本库中敏感数据单元,建立金融隐私模型并进行计量。[结果/结论]计量结果表明,金融业机构制作的个人数据(个人账户数据、金融交易数据、个人借贷数据)的隐私性较高,金融业调用的个人数据(财产数据、司法数据、医疗数据)的隐私性次之,金融业采集的个人身份数据(身份基本数据、身份鉴定数据)的隐私性较低。
[Purpose/significance] In the current financial data industry standard specification,personal financial data privacy is divided into three or four levels according to the “degree of impact caused by damage or leakage”,but all of them lack quantitative research support.This paper mines financial privacy related texts,and quantitatively calculates the privacy value of financial data items,which provides a scientific and objective basis for personal financial privacy classification.[Method/process] Survey financial privacy related literature,establish financial privacy text base;according to the financial data industry standards and specifications,the financial privacy glossary framework is established to collect financial sensitive words;investigate financial institutions,collect their customer personal data,and complete financial sensitive vocabulary;based on the syntactic structure of the text,financial privacy verbs and financial sensitivity words are separated from the text base,and a financial privacy thesaurus is established;based on the privacy vocabulary,the sensitive data units are extracted from the financial text database,and the financial privacy model is established and measured.[Result/conclusion] The measurement results show that the privacy of personal data produced by financial institutions(personal account data,financial transaction data,personal lending data) is higher,the privacy of personal data used by the financial industry(property data,judicial data,medical data) is the next,and the privacy of personal identity data collected by the financial industry(identity basic data,identity identification data) is low.
出处
《情报理论与实践》
CSSCI
北大核心
2023年第9期105-114,86,共11页
Information Studies:Theory & Application
基金
国家社会科学基金重大项目“政府数据的隐私风险计量与保护机制创新研究”的成果之一,项目编号:21&ZD338。
关键词
个人金融数据
数据隐私
隐私计量
金融隐私词表
personal financial data
data privacy
privacy measurement
financial privacy vocabulary
