A0195
Title: Data-adaptive automatic threshold calibration for stability selection
Authors: Martin Huang - University of Sydney (Australia) [presenting]
Samuel Muller - University of Sydney (Australia)
Garth Tarr - University of Sydney (Australia)
Abstract: Stability selection has become a popular method for enhancing the performance of variable selection algorithms owing to its false discovery rate properties. However, attaining these favorable characteristics depends on properly specifying the stable threshold parameter, which can be challenging at times. Arbitrary choices of this parameter can result in substantially different outcomes, as the variable selection probabilities are dataset-dependent. In response to this challenge, a data-adaptive automatic threshold selection algorithm is proposed that streamlines stability selection through the simplification of the threshold specification process.
