Table 3 |
||||||||||
|
Two medoid classification results for the same dataset |
||||||||||
|
Training Set |
Testing Set |
|||||||||
|
Cases |
Controls |
Cases |
Controls |
|||||||
|
Result |
Medoid |
Corr |
Incorr |
Medoid |
Corr |
Incorr |
Corr |
Incorr |
Corr |
Incorr |
|
|
||||||||||
|
1 |
35 |
4 |
1 |
34 |
5 |
1 |
19 |
1 |
19 |
1 |
|
2 |
36 |
3 |
1 |
36 |
3 |
1 |
19 |
1 |
19 |
1 |
|
|
||||||||||
|
These results are for the same dataset as in Table 2 after a single mixing of the order of Cases and Controls after a hypothetical division of the samples into a training set and a testing set. The first classifier uses features 79, 114, 135, 137, 140, 224 and 300, and the second uses features 62, 65, 141, 146, 156, 211, and 229. |
||||||||||
|
Luke and Collins BMC Bioinformatics 2008 9:545 doi:10.1186/1471-2105-9-545 |
||||||||||