Abstract: Data Allocation Strategy(DAS) system shows its advantages to address data leakage problems. This paper further illustrates the recent advances in DAS-based data leakage detection techniques including the first and second generation detection methods. Guilt model is demonstrated the potential as a suitable candidate for data leakage detection problems. The algorithms on explicit data request,random object allocation and agent selection are analyzed and evaluated with supporting data. Amodification on data allocation design is proposed. The experiment shows that the proposed algorithvs increases the possibility of detecting the guilt agents which leak the data. Consequently,the future works for guilt model are discussed.

Key words: Data Allocation Strategy(DAS), guilt model, Data Leakage Protection(DLP), data privacy, pseudo data

摘要： 使用数据分配策略(DAS)能有效检测数据泄露。鉴于此,分别介绍基于第1 代和第2 代DAS 的检测方法。利用过失模型在解决数据泄漏检测问题上的优势,研究基于过失代理模型的数据分配算法,对其中的明确数据请求分配算法、随机假对象分配算法和优化Agent 选择算法进行优化,并对代理模型进行实验仿真,结果表明,优化算法在添加伪数据后能明显提高对过失代理的检测率。此外,还给出针对过失模型的研究方向。

关键词: 数据分配策略, 过失模型, 数据泄漏保护, 数据隐私, 伪数据

CLC Number: 

• TP309