摘要
现有研究表明,社交网络发布数据的隐私和效用之间的权衡成为一个重要和具有挑战性的问题,同时用户的敏感属性是社交网络中用户的重要隐私信息.为了高效解决社交网络中用户的敏感属性泄露问题,提出一种基于效用的用户属性个性化保护算法PKD-U.首先计算属性的敏感度,将保护对象由敏感属性精确到个别的敏感属性值;然后,将含有敏感属性值的非关键节点进行分割或合并,并匿名关键节点属性,从而生成具有隐私能力的匿名社交网络.实验结果表明,相比K-degree算法,该算法在有效抵制用户敏感信息泄露的情况下能更大限度地保证数据具有较高可用性.
Existing studies have shown that the weigh between privacy and utility upon anonymous social network publishing data becomes an important and challenging issue, while users' sensitive attributes are important privacy information. In order to efficiently solve the sensitive attributes leaking problem in social network, we propose a personalized protection algorithm based utility PKD-U model. First, calculate the sensitivity of the user attributes, the protection objects change from a sensitive attribute to specific attribute values;then, partition or consolidate the non-critical nodes containing sensitive attribute value, and generalize attributes of the key nodes. Experimental results show that, compared to K-degree algorithm, PKD-U algorithm can guarantee higher data availability, while effectively against private information disclosure.
出处
《小型微型计算机系统》
CSCD
北大核心
2017年第7期1490-1494,共5页
Journal of Chinese Computer Systems
基金
国家社会科学基金项目(16CJY056)资助
关键词
社交网络
隐私保护
个性化
匿名
数据效用
social networking
privacy preserving
personalized
anonymity
data utility