Research Article
Probing for Sparse and Fast Variable Selection with Model-Based Boosting
Figure 1
True positive rate (on -axis) and false discovery rate (on -axis) for three different, boosting-based variable selection algorithms, probing (black), stability selection (green), cross-validation (blue), and different simulation settings: , , and . All settings of stability selection are combined. Shaded areas are smooth hulls around all observed values.