Aggregation of Multiple Knockoffs
This is an incremental improvement for researchers in statistical inference and variable selection, addressing a known bottleneck in stability.
The paper tackles the instability of the Knockoff Inference procedure by introducing Aggregation of Multiple Knockoffs (AKO), which improves stability and power while maintaining False Discovery Rate control, as demonstrated in experiments on synthetic and real datasets.
We develop an extension of the Knockoff Inference procedure, introduced by Barber and Candes (2015). This new method, called Aggregation of Multiple Knockoffs (AKO), addresses the instability inherent to the random nature of Knockoff-based inference. Specifically, AKO improves both the stability and power compared with the original Knockoff algorithm while still maintaining guarantees for False Discovery Rate control. We provide a new inference procedure, prove its core properties, and demonstrate its benefits in a set of experiments on synthetic and real datasets.