计算机工程与设计
計算機工程與設計
계산궤공정여설계
COMPUTER ENGINEERING AND DESIGN
2010年
7期
1444-1446
,共3页
数据挖掘%隐私保护%聚类%主成份分析%Laplace噪音
數據挖掘%隱私保護%聚類%主成份分析%Laplace譟音
수거알굴%은사보호%취류%주성빈분석%Laplace조음
data mining%privacy preserving%clustering%principal analysis%Laplace noise
为了解决在极端情况下数据挖掘中隐私泄露的问题,分析了在数据聚类时增加Laplace噪音可以避免隐私泄露的原理,结合主成份分析与噪音扰动方法,提出了一种限制隐私泄露的隐私保护聚类算法.该算法利用主成份分析除掉了数据的相关性,将Laplace噪音加入数据的主成份向量中,然后计算被扰动的数据之间距离变化值,这样可以避免扰动后的数据被还原,以达到在隐私保护聚类挖掘中限制隐私泄露的目的.仿真实验结果表明,该算法对于数据聚类时限制隐私泄露是正确有效的.
為瞭解決在極耑情況下數據挖掘中隱私洩露的問題,分析瞭在數據聚類時增加Laplace譟音可以避免隱私洩露的原理,結閤主成份分析與譟音擾動方法,提齣瞭一種限製隱私洩露的隱私保護聚類算法.該算法利用主成份分析除掉瞭數據的相關性,將Laplace譟音加入數據的主成份嚮量中,然後計算被擾動的數據之間距離變化值,這樣可以避免擾動後的數據被還原,以達到在隱私保護聚類挖掘中限製隱私洩露的目的.倣真實驗結果錶明,該算法對于數據聚類時限製隱私洩露是正確有效的.
위료해결재겁단정황하수거알굴중은사설로적문제,분석료재수거취류시증가Laplace조음가이피면은사설로적원리,결합주성빈분석여조음우동방법,제출료일충한제은사설로적은사보호취류산법.해산법이용주성빈분석제도료수거적상관성,장Laplace조음가입수거적주성빈향량중,연후계산피우동적수거지간거리변화치,저양가이피면우동후적수거피환원,이체도재은사보호취류알굴중한제은사설로적목적.방진실험결과표명,해산법대우수거취류시한제은사설로시정학유효적.
To deal with privacy breaches in the extreme cases of data mining, the theory that the privacy breaches could have avoided are analyzed, when the Laplace noise are added to the data clustering. Combining the methods of the principal analysis and the Laplace noise, a privacy preserving clustering algorithm of limiting privacy breaches is proposed. The algorithm got rid of the relevance of the data to use principal component analysis, and the distance of the disturbance chance is calculated after added Laplace noise into the main component of the data vector, those made the disturbed data not recover and played a part in the limiting privacy breaches in the privacy preserving clustering algorithm. Results of the simulation experiment indicated the proposed algorithm for limiting privacy breaches in the data clustering is correct and effective.
