Privacy-preserving attribute reduction algorithm based on relative granularity over horizontally partitioned multi-decision tables
水平划分多决策表下基于相对粒度的隐私保护属性约简算法*

Aiming at the horizontally partitioned multi-decision tables,this paper proposed an algorithm based on relative granularity for privacy-preserving distributed attribute reduction,which could solve the problem that multiple parties carried out attribute reduction computation in distributed environment without sharing private data.The algorithm could compute global attribute reduction based on the attribute reduction idea of relative granularity,and used semi-trusted third party and secure multi-party technology to design a privacy-preserving protocol for computing global relative granularity,which could get accurate attribute reduction effect in the premise of no sharing of private information among participators.Analysis results show the proposed algorithm is effective and efficient.