Table 1.
Database name | Original dataset | Encrypted dataset | ||||
|
Tree sizea, n | Leavesb, n | Accuracyc, % | Tree sizea, n | Leavesb, n | Accuracyd, % |
Abalone | 2312 | 1183 | 21.97 | 2312 | 1183 | 21.97 |
Acute inflammations | 5 | 3 | 100.00 | 5 | 3 | 100.00 |
Arrhythmia | 99 | 50 | 71.43 | 99 | 50 | 71.43 |
Audiology (standardized) | 54 | 32 | 83.12 | 54 | 32 | 83.12 |
Breast cancer | 6 | 4 | 68.04 | 6 | 4 | 68.04 |
Breast cancer Wisconsin (original) | 27 | 14 | 95.38 | 27 | 14 | 95.38 |
Breast tissue | 29 | 15 | 47.22 | 29 | 15 | 47.22 |
Cardiotocography | 19 | 14 | 98.34 | 33 | 25 | 98.34 |
Contraceptive method choice | 263 | 157 | 55.29 | 263 | 157 | 55.29 |
Covertype | 29,793 | 14,897 | 93.59 | 29,793 | 14,897 | 93.59 |
Dermatology | 41 | 31 | 92.74 | 41 | 31 | 92.74 |
Echocardiogram | 9 | 5 | 70.37 | 9 | 5 | 70.37 |
Ecoli | 43 | 22 | 78.95 | 43 | 22 | 78.95 |
Haberman’s survival | 5 | 3 | 75.96 | 5 | 3 | 75.96 |
Hepatitis | 21 | 11 | 79.25 | 21 | 11 | 79.25 |
Horse colic | 29 | 18 | 68.55 | 29 | 18 | 68.55 |
Iris | 9 | 5 | 96.08 | 9 | 5 | 96.08 |
Lung cancer | 16 | 10 | 63.64 | 16 | 10 | 63.64 |
Lymphography | 34 | 21 | 78.00 | 34 | 21 | 78.00 |
Mammographic mass | 15 | 12 | 82.26 | 15 | 12 | 82.26 |
Mushroom | 30 | 25 | 100.00 | 30 | 25 | 100.00 |
Pima Indians diabetes | 39 | 20 | 76.25 | 39 | 20 | 76.25 |
Post-operative patient | 1 | 1 | 70.97 | 1 | 1 | 70.97 |
Primary tumor | 88 | 47 | 39.13 | 88 | 47 | 39.13 |
Seeds | 15 | 8 | 97.18 | 15 | 8 | 97.18 |
Soybean (large) | 93 | 61 | 90.52 | 93 | 61 | 90.52 |
Spectf heart | 17 | 9 | 66.67 | 17 | 9 | 66.67 |
Statlog (heart) | 45 | 27 | 76.09 | 45 | 27 | 76.09 |
Yeast | 369 | 185 | 58.81 | 369 | 185 | 58.81 |
Zoo | 17 | 9 | 94.12 | 17 | 9 | 94.12 |
aNumber of nodes (measurements) in a tree.
bNumber of decision rules in a tree.
cPercentage of correctly classified original items with respect to all items (ie, the number of times the tree's rules lead to the right decision).
dPercentage of correctly classified encrypted items with respect to all items (ie, the number of times the tree's rules lead to the right decision).