Research Article

[Retracted] Differentially Private Singular Value Decomposition for Training Support Vector Machines

Table 4

Performance comparison of algorithms on different datasets.

DatasetsɛAlgorithmAccuracySV
MeanStdMaxMinMeanStdMaxMin

A1a--SVM83.49------754------
0.1DPSVD83.350.1583.6183.247636771756
AG83.080.2483.3082.687434747738
DPPCA-SVM82.580.5883.3081.9377112783754
0.5DPSVD83.500.2283.8683.307559768743
AG83.090.1683.1882.806863689682
DPPCA-SVM83.070.5083.4982.2476414778747
1DPSVD83.590.1383.7483.437535759748
AG83.450.2483.8083.186974702693
DPPCA-SVM83.460.4184.1182.9975614778742

Mushrooms--SVM99.90------617------
0.1DPSVD99.890.0199.9099.8863939700607
AG99.180.0299.2199.155183521514
DPPCA-SVM99.490.4299.8999.0268367747604
0.5DPSVD99.900.0199.9099.8963325674607
AG99.190.0599.2699.145245531517
DPPCA-SVM99.590.3899.9099.0676350811687
1DPSVD99.900.0099.9099.9060226625559
AG99.790.0499.8399.7344522469417
DPPCA-SVM99.830.0899.9899.7865181779559

Musk--SVM93.95------1351------
0.1DPSVD94.080.1194.2393.9513511513691330
AG88.960.0889.0988.8918651018761855
DPPCA-SVM93.970.2094.2393.741379813911369
0.5DPSVD94.140.1194.2794.0013591413791341
AG88.940.0188.9588.921866618741858
DPPCA-SVM94.100.1594.3593.9713361513551315
1DPSVD94.140.1094.2994.0413451013581333
AG88.930.0288.9588.9118721018871860
DPPCA-SVM94.190.1794.3593.9213184013841275

Splice--SVM94.30------607------
0.1DPSVD91.080.7592.4090.6063517662616
AG90.560.8391.3089.3059116605568
DPPCA-SVM87.140.3287.4086.7064316660619
0.5DPSVD92.000.8092.8090.7061010625600
AG93.580.3894.0093.005885595582
DPPCA-SVM87.220.7388.4086.6065934706615
1DPSVD92.360.5693.1091.8061819645594
AG93.560.3193.9093.105943599591
DPPCA-SVM87.360.9488.3086.0064121667614