gdp
所属分类：matlab编程
开发工具：matlab
文件大小：688KB
下载次数：62
上传日期：2014-03-29 02:08:52
上 传 者：
lifebud
说明： 关于数据挖掘与隐私保护的源代码及论文，主要用到k-means的聚类分析
(About Data Mining and Privacy Protection of the source code and papers, mainly used k-means clustering analysis)
文件列表:
gdp (0, 2014-03-04)
gdp\Geometric Data Perturbation for Privacy Preserving Outsourced Data Mining.pdf (839357, 2014-03-03)
gdp\compareSamplePDF.m (1687, 2012-06-22)
gdp\gdp.m (986, 2012-10-06)
gdp\gdp_data.m (195, 2012-09-29)
gdp\ica (0, 2014-03-03)
gdp\ica\Contents.m (1296, 2004-07-28)
gdp\ica\demosig.m (748, 2003-04-06)
gdp\ica\dispsig.m (402, 2003-04-06)
gdp\ica\fastica.m (18420, 2004-07-28)
gdp\ica\fasticag.m (19214, 2004-07-28)
gdp\ica\fpica.asv (26006, 2006-06-27)
gdp\ica\fpica.m (25976, 2006-06-27)
gdp\ica\gui_adv.m (13126, 2004-07-28)
gdp\ica\gui_advc.m (7411, 2003-09-09)
gdp\ica\gui_cb.m (19416, 2003-09-11)
gdp\ica\gui_cg.m (2704, 2003-04-06)
gdp\ica\gui_help.m (14533, 2004-07-28)
gdp\ica\gui_l.m (5129, 2004-07-28)
gdp\ica\gui_lc.m (3665, 2003-09-12)
gdp\ica\gui_s.m (5017, 2004-07-28)
gdp\ica\gui_sc.m (2402, 2003-09-09)
gdp\ica\icaplot.m (13259, 2003-04-06)
gdp\ica\pcamat.m (12075, 2003-12-16)
gdp\ica\remmean.m (461, 2003-04-06)
gdp\ica\whitenv.m (2842, 2003-10-13)
gdp\iris_nolabel (2400, 2012-06-22)
gdp\maxRot.m (920, 2012-03-25)
gdp\normalization.m (279, 2012-03-25)
gdp\rp.m (154, 2014-03-22)
gdp\test_gdp.m (195, 2012-09-29)
gdp\test_ica.m (1145, 2012-10-06)
gdp\test_io_attack.m (1152, 2012-10-06)
gdp\test_rp.m (207, 2012-10-06)
the ica directory contains the FastICA code from http://research.ics.tkk.fi/ica/fastica/
maxRot.m - return a rotation matrix that maximizes the resilience to naive estimation
test_ica.m - test the resilience of perturbed data to the ICA attack
test_io_attack.m - test the resilience of perturbed data to the known input/output attack
normalization.m - normalize each dimension to mean=0, standard deviation =1
compareSamplePDF.m - the ica attack depends on the match of column PDFs to identify the pairs of input/ouput columns, using sampling methods to approximately matach
gdp.m - input the original matrix and the std of random noise, return the privacy guarantees in terms of the three types of attacks and the perturbation parameters: rotation matrix Rt and the translation vector tr
rp.m - random projection. The only difference from gdp is the elements of perturbation matrix are drawn from N(0,1)
privacy guarantee is defined as RMSE(x, hat(x)), where x and hat(x) are normalized to mean=0, and std =1. You may need to convert it to be relative to the domain
Note that the resilience to the IO attack is determined by the noise std.
test_gdp.m and test_rp.m uses the sample data iris_nolabel.
近期下载者：
相关文件：
收藏者：