基于频繁项集发现的匿名隐私保护算法

摘要

通过定义考虑权重的匿名表效用度量函数,用于在泛化步骤决定下一个泛化路径以取得较好的泛化效果,在此基础上提出利用频繁项集发现思想的微观数据表匿名隐私保护算法ABFI( algorithm based on frequent set mining),匿名过程仅仅对不满足隐私保护要求等价组中准码属性取值进行泛化.实验结果表明,该方法可以减少信息损失,求解得到更加符合数据分析任务需求的局部最优匿名表.%Taking into consider data distribution, this paper proposed ABFI. ABFI defined a function referred to weight to make which quasi attribution would be generalized. Experiments show that RTBAGL obtains a higher increase in data quality.