Privacy preserving classification based on randomization and reconstruction

被引：0

作者：

Zhang, Peng ^{[1
]}

Tong, Yunhai ^{[1
]}

Tang, Shiwei ^{[1
]}

Yang, Dongqing ^{[1
]}

机构：

[1] Peking Univ, Sch Elect Engn & Comp Sci, Beijing 100871, Peoples R China

来源：

DYNAMICS OF CONTINUOUS DISCRETE AND IMPULSIVE SYSTEMS-SERIES B-APPLICATIONS & ALGORITHMS | 2007年 / 14卷

关键词：

data mining; privacy preservation; classification; Naive Bayes; data randomization; distribution reconstruction;

D O I：

暂无

中图分类号：

O29 [应用数学];

学科分类号：

070104 ;

摘要：

Privacy preserving classification is to develop a classifier without precise access to the original data. In order to improve the applicability with higher privacy and better accuracy, we present a novel Privacy Preserving Naive Bayes (PPNB) classification method that consists of two steps: first, the original data set is distorted by a new. randomization approach; second, a naive Bayes classifier is implemented on the distorted data set to predict the class labels for unknown samples. Besides being analyzed in applicability, privacy, accuracy, and efficiency, the effectiveness of our PPNB classification method is also validated by the experiments.

引用

页码：166 / 173

页数：8

共 10 条

[1]

Agrawal D., 2001, Proceedings of the 20th ACM SIGACT-SIGMOD-SIGART Symposium on Principles of Database Systems, P247, DOI DOI 10.1145/375551.375602

[2]

Agrawal S, 2004, LECT NOTES COMPUT SC, V2973, P113

[3]

[Anonymous], 2000, Privacy-preserving data mining, DOI DOI 10.1145/342009.335438

[4]

Clifton C., 2000, Journal of Computer Security, V8, P281

[5]

Clifton C, 2003, P IEEE ICDM WORKSH P, P3

[6]

Clifton C., 2002, SIGKDD EXPLOR NEWSLE, V4, P28, DOI [10.1145/772862.772867, DOI 10.1145/772862.772867]

[7]

DU W, 2003, P 9 ACM SIGKDD INT C, P505

[8]

Johnsten T., 2002, P IEEE INT C PRIVAC, P9

[9]

Rizvi S. J., 2002, Proceedings of the Twenty-eighth International Conference on Very Large Data Bases, P682

[10]

VAIDYA J, 2004, P 4 SIAM INT C DAT M

← 1 →