We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

stat

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Mathematics > Statistics Theory

Title: Trimming Stability Selection increases variable selection robustness

Authors: Tino Werner
Abstract: Contamination can severely distort an estimator unless the estimation procedure is suitably robust. This is a well-known issue and has been addressed in Robust Statistics, however, the relation of contamination and distorted variable selection has been rarely considered in literature. As for variable selection, many methods for sparse model selection have been proposed, including Stability Selection which is a meta-algorithm based on some variable selection algorithm in order to immunize against particular data configurations. We introduce the variable selection breakdown point that quantifies the number of cases resp. cells that have to be contaminated in order to let no relevant variable be detected. We show that particular outlier configurations can completely mislead model selection and argue why even cell-wise robust methods cannot fix this problem. We combine the variable selection breakdown point with resampling, resulting in the Stability Selection breakdown point that quantifies the robustness of Stability Selection. We propose a trimmed Stability Selection which only aggregates the models with the lowest in-sample losses so that, heuristically, models computed on heavily contaminated resamples should be trimmed away. We provide a short simulation study that reveals both the potential of our approach as well as the fragility of variable selection, even for an extremely small cell-wise contamination rate.
Subjects: Statistics Theory (math.ST); Methodology (stat.ME)
Cite as: arXiv:2111.11818 [math.ST]
  (or arXiv:2111.11818v1 [math.ST] for this version)

Submission history

From: Tino Werner [view email]
[v1] Tue, 23 Nov 2021 12:17:10 GMT (72kb,D)

Link back to: arXiv, form interface, contact.