Creates a new SDCA learner for LogisticLoss logistic-regression. Pure L2 or L1 regularization can be obtained using the #setAlpha(double) alpha parameter.
()
| 84 | * {@link #setAlpha(double) alpha} parameter. |
| 85 | */ |
| 86 | public SDCA() |
| 87 | { |
| 88 | this(1e-5); |
| 89 | } |
| 90 | |
| 91 | /** |
| 92 | * <br>The implementation will use Elastic-Net regularization by default. |