Table 3 Prediction error of support vector machine models with different width of radial kernel. The baseline width is the inverse of the dimension of the data (in this case Baseline will be 0.25)
From: Supervised learning methods in modeling of CD4+ T cell heterogeneity
Width of radial | IL17 | RORgt | IFNγ | Tbet | FOXP3 | Sum of prediction error |
|---|---|---|---|---|---|---|
Baseline | 0.181 | 0.179 | 0.146 | 0.122 | 0.0355 | 0.665 |
1 | 0.189 | 0.192 | 0.149 | 0.126 | 0.0349 | 0.691 |
0.1 | 0.193 | 0.192 | 0.160 | 0.130 | 0.0360 | 0.711 |
0.01 | 0.257 | 0.263 | 0.216 | 0.148 | 0.0366 | 0.920 |
0.001 | 0.343 | 0.351 | 0.259 | 0.174 | 0.0368 | 1.163 |