Table 1 Testing accuracies associated with the highest training accuracies under different feature selections for the MICC data set.
From: Supervised learning-based tagSNP selection for genome-wide disease classifications
Feature Selection | Testing accuracy (mean value ± standard deviation, %) | |||
|---|---|---|---|---|
NMSC | NBC | SVM | UDC | Â |
MSW-MSC | 76.0 ± 3.4 | 75.1 ± 3.0 | 73.1 ± 4.5 | 73.6 ± 2.9 |
MMW-MSC | 77.4 ± 2.9 | 75.9 ± 3.0 | 74.4 ± 2.3 | 74.8 ± 4.6 |
NBC-MSC | 75.1 ± 3.1 | 73.2 ± 2.4 | 74.2 ± 4.1 | 75.2 ± 2.6 |
NMSC-MSC | 75.0 ± 4.5 | 75.0 ± 2.9 | 74.0 ± 3.7 | 72.7 ± 3.9 |
DENFIS-MSC | 76.9± 3.2 | 74.2 ± 3.4 | 74.9 ± 4.4 | 75.6 ± 2.8 |
SVMRFE | 77.0 ± 4.2 | 73.9 ± 2.7 | 73.1 ± 4.0 | 74.4 ± 3.2 |
T-TEST | 75.6 ± 2.6 | 76.4 ± 3.0 | 74.5 ± 3.1 | 75.9 ± 3.6 |
LOGICFS | 54.4±1.5 | |||