Table 5. Mean classification accuracies (%) of OSWLDA, OPCALDA, and OLDA evaluated on data set A using cross-validation.
Algorithms | Intensification sequences | ||||||||||||||||
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | |||
OSWLDA | 45 | 1 | 40.8 | 62.6 | 75.8 | 81.6 | 85.0 | 88.0 | 90.6 | 92.6 | 92.2 | 93.6 | 94.0 | 93.6 | 93.4 | 94.0 | 94.6 |
5 | 47.4 | 67.0 | 79.0 | 83.8 | 88.0 | 89.4 | 91.6 | 94.0 | 94.6 | 94.2 | 94.6 | 94.8 | 95.4 | 95.6 | 96.0 | ||
10 | 47.2 | 68.2 | 80.4 | 86.2 | 90.4 | 90.4 | 93.2 | 94.8 | 95.4 | 95.4 | 96.2 | 96.0 | 97.0 | 96.4 | 96.8 | ||
15 | 47.8 | 68.6 | 80.6 | 85.4 | 91.4 | 91.6 | 94.0 | 95.4 | 96.2 | 96.4 | 97.0 | 96.6 | 97.2 | 97.8 | 97.8 | ||
20 | 48.4 | 68.0 | 81.2 | 86.4 | 91.8 | 92.8 | 94.2 | 96.2 | 96.0 | 96.6 | 97.2 | 97.2 | 97.8 | 97.6 | 98.2 | ||
25 | 47.8 | 69.0 | 81.0 | 86.2 | 91.6 | 92.8 | 94.6 | 96.6 | 96.2 | 97.0 | 97.8 | 97.4 | 98.4 | 98.6 | 98.2 | ||
30 | 47.8 | 69.2 | 80.8 | 85.6 | 91.0 | 93.0 | 94.8 | 96.4 | 97.0 | 96.8 | 98.2 | 97.6 | 98.4 | 98.6 | 98.8 | ||
35 | 47.4 | 68.8 | 80.0 | 85.8 | 91.2 | 92.8 | 94.4 | 96.4 | 97.2 | 97.4 | 98.6 | 98.0 | 98.4 | 98.6 | 98.8 | ||
40 | 46.2 | 68.0 | 79.6 | 85.0 | 91.2 | 92.2 | 94.8 | 97.0 | 97.4 | 97.2 | 98.4 | 98.4 | 98.4 | 98.6 | 98.8 | ||
45 | 43.4 | 68.2 | 80.2 | 86.0 | 90.2 | 93.4 | 95.4 | 96.4 | 97.4 | 97.2 | 98.2 | 98.2 | 98.4 | 98.6 | 99.0 | ||
OPCALDA | 45 | 1 | 46.0 | 63.8 | 76.6 | 83.2 | 87.8 | 88.8 | 91.2 | 92.8 | 92.6 | 92.8 | 93.6 | 94.2 | 93.8 | 94.0 | 94.2 |
5 | 46.0 | 68.8 | 78.8 | 83.8 | 89.2 | 90.0 | 91.2 | 93.2 | 94.2 | 94.2 | 95.4 | 95.0 | 94.6 | 94.4 | 95.4 | ||
10 | 46.4 | 67.6 | 79.2 | 84.6 | 90.0 | 90.0 | 92.2 | 93.4 | 95.6 | 95.6 | 96.0 | 96.0 | 95.8 | 96.4 | 96.8 | ||
15 | 47.4 | 68.2 | 80.0 | 86.0 | 90.6 | 92.0 | 93.8 | 95.2 | 96.2 | 96.4 | 96.8 | 96.8 | 96.8 | 97.2 | 98.0 | ||
20 | 47.0 | 68.2 | 80.2 | 85.4 | 90.4 | 93.2 | 94.2 | 95.6 | 96.4 | 96.6 | 97.0 | 97.0 | 97.4 | 97.8 | 98.0 | ||
25 | 47.0 | 66.6 | 80.2 | 85.4 | 91.0 | 92.4 | 93.8 | 95.6 | 96.6 | 97.0 | 97.4 | 97.2 | 97.8 | 98.2 | 98.2 | ||
30 | 47.4 | 67.0 | 80.6 | 85.4 | 91.0 | 92.2 | 94.4 | 95.4 | 97.0 | 96.8 | 97.6 | 97.4 | 98.0 | 98.2 | 98.6 | ||
35 | 47.0 | 67.4 | 80.2 | 85.4 | 91.2 | 92.0 | 94.0 | 95.6 | 97.2 | 97.2 | 98.0 | 97.8 | 98.0 | 98.2 | 98.6 | ||
40 | 46.6 | 67.0 | 80.6 | 85.4 | 91.4 | 92.2 | 94.2 | 96.0 | 97.6 | 97.4 | 98.4 | 98.0 | 98.0 | 98.2 | 98.6 | ||
45 | 46.8 | 67.2 | 80.4 | 86.0 | 91.4 | 91.8 | 94.2 | 96.2 | 97.4 | 97.2 | 98.4 | 98.0 | 98.0 | 98.2 | 98.6 | ||
OLDA | 45 | 1 | 3.8 | 3.4 | 4.4 | 3.8 | 2.6 | 2.8 | 3.4 | 2.8 | 2.6 | 2.6 | 2.8 | 3.6 | 4.0 | 4.2 | 4.2 |
5 | 46.6 | 67.2 | 79.0 | 84.4 | 89.4 | 90.4 | 92.2 | 93.6 | 94.6 | 94.6 | 95.4 | 95.6 | 95.2 | 95.6 | 96.2 | ||
10 | 47.8 | 68.4 | 80.0 | 84.2 | 89.8 | 90.6 | 93.2 | 94.6 | 95.2 | 95.6 | 96.8 | 96.8 | 96.6 | 97.0 | 97.4 | ||
15 | 47.6 | 68.6 | 81.8 | 85.4 | 90.6 | 91.8 | 94.2 | 95.2 | 96.0 | 96.2 | 96.8 | 96.8 | 97.8 | 97.8 | 98.6 | ||
20 | 47.2 | 68.6 | 81.8 | 85.6 | 90.8 | 92.0 | 94.4 | 96.0 | 96.4 | 96.8 | 97.8 | 97.8 | 98.4 | 98.8 | 99.2 | ||
25 | 46.8 | 68.2 | 82.0 | 86.2 | 91.2 | 92.2 | 94.8 | 96.4 | 96.6 | 97.0 | 98.0 | 97.8 | 98.6 | 98.8 | 99.6 | ||
30 | 46.4 | 67.4 | 81.8 | 86.4 | 91.0 | 93.0 | 94.6 | 96.6 | 96.8 | 97.2 | 98.2 | 98.2 | 98.6 | 98.8 | 99.4 | ||
35 | 46.4 | 68.0 | 81.6 | 86.8 | 91.0 | 93.0 | 94.6 | 96.2 | 97.0 | 97.6 | 98.8 | 98.4 | 98.6 | 98.8 | 99.4 | ||
40 | 46.0 | 68.0 | 81.4 | 86.8 | 90.8 | 93.0 | 94.6 | 96.4 | 97.2 | 97.6 | 98.8 | 98.4 | 98.6 | 98.6 | 99.4 | ||
45 | 46.0 | 68.0 | 81.0 | 86.8 | 90.6 | 93.0 | 94.6 | 96.4 | 97.2 | 97.6 | 98.8 | 98.6 | 98.6 | 98.6 | 99.4 |
The best accuracy among all for each algorithm and each repetition is written in bold and the worst is underlined. An overlapped ensemble classifier becomes an ensemble classifier with naive partitioning when and . The classifier is equivalent to a single classifier when and .