抗恶意敌手的保密关联挖掘方案

Secure association mining scheme against malicious adversaries

为防止在数据挖掘中泄漏隐私数据,利用安全多方计算在数据挖掘中对隐私数据进行保密计算,得到正确结果,保证隐私数据的安全。大多数的安全多方计算模型是半诚实模型,而在现实生活中难免存在恶意行为,分析现有保密关联挖掘方案中可能存在的恶意行为,基于零知识证明和分割-选择等方法,设计抗恶意敌手的保密关联挖掘方案。通过与现有方案对比,验证了协议可有效避免或发现恶意行为,具有较高的安全性和实用性。

To prevent the leakage of private data in data mining,secure multi-party computing was used to calculate the privacy data in data mining,which not only got the correct results,but ensured the security of the privacy data.Most secure multi-party computing models are semi-honest models,but malicious behaviors are unavoidable in real life.The possible malicious behaviors in existing secret association mining protocol were analyzed,and a secret association mining protocol against malicious adversaries based on zero-knowledge proof and cut-choose method was designed.Compared with the existing schemes,it is verified that the protocol can effectively avoid or detect malicious acts,and has high security and practicability.