Table 9 The average recall of the 3NN classifier in the multi-manifold and single manifold approaches
Dataset | PCA | LPP | NPE | Multi-Manifold |
|---|---|---|---|---|
ecoli1 | 0.89 ± 0.139 (4) | 0.93 ± 0.130 (1) | 0.93 ± 0.130 (1) | 0.93 ± 0.130 (1) |
ecoli2 | 0.95 ± 0.110 (1) | 0.95 ± 0.110 (1) | 0.95 ± 0.110 (1) | 0.95 ± 0.111 (1) |
ecoli3 | 0.90 ± 0.229 (2) | 0.85 ± 0.229 (3) | 0.85 ± 0.229 (3) | 0.93 ± 0.225 (1) |
ecoli4 | 0.90 ± 0.200 (4) | 0.95 ± 0.150 (1) | 0.95 ± 0.150 (1) | 0.95 ± 0.150 (1) |
ecoli0147vs56 | 0.85 ± 0.189 (3) | 0.85 ± 0.189 (3) | 0.88 ± 0.183 (1) | 0.88 ± 0.183 (1) |
ecoli034_5 | 0.90 ± 0.200 (1) | 0.90 ± 0.200 (1) | 0.90 ± 0.200 (1) | 0.90 ± 0.200 (1) |
ecoli0147_2356 | 0.82 ± 0.240 (1) | 0.82 ± 0.240 (1) | 0.82 ± 0.240 (1) | 0.82 ± 0.240 (1) |
glass0 | 0.89 ± 0.106 (1) | 0.89 ± 0.124 (1) | 0.87 ± 0.134 (4) | 0.89 ± 0.124 (1) |
glass0123456 | 0.96 ± 0.080 (1) | 0.96 ± 0.080 (1) | 0.96 ± 0.080 (1) | 0.96 ± 0.80 (1) |
kddcup-buffer_overflow_vs_back | 1 ± 0 (1) | 1 ± 0 (1) | 1 ± 0 (1) | 1 ± 0 (1) |
new-thyroid1 | 0.93 ± 0.200 (1) | 0.93 ± 0.200 (1) | 0.93 ± 0.200 (1) | 0.93 ± 0.200 (1) |
page-blocks-1-3_vs_4 | 0.93 ± 0.133 (1) | 0.93 ± 0.133 (1) | 0.93 ± 0.133 (1) | 0.93 ± 0.133 (1) |
Pima | 0.63 ± 0.167 (1) | 0.62 ± 0.177 (2) | 0.61 ± 0.153 (4) | 0.62 ± 0.208 (2) |
segment0 | 0.99 ± 0.027 (1) | 0.99 ± 0.020 (1) | 0.99 ± 0.020 (1) | 0.99 ± 0.027 (1) |
shuttle_2_vs_5 | 1 ± 0 (1) | 1 ± 0 (1) | 1 ± 0 (1) | 1 ± 0 (1) |
vehicle2–1 | 0.96 ± 0.041 (2) | 0.95 ± 0.061 (3) | 0.95 ± 0.061 (3) | 0.97 ± 0.042 (1) |
vowel0 | 0.88 ± 0.245 (4) | 0.90 ± 0.213 (1) | 0.90 ± 0.213 (1) | 0.90 ± 0.213 (1) |
Wisconsin | 0.98 ± 0.027 (1) | 0.98 ± 0.028 (1) | 0.98 ± 0.028 (1) | 0.98 ± 0.027 (1) |
Average Ratings | 1.72 (4) | 1.39 (2) | 1.56 (3) | 1.06 (1) |
Average Recall | 0.91 (2) | 0.91 (2) | 0.91 (2) | 0.92 (1) |