The Power of Batching in Multiple Hypothesis Testing

Tijana Zrnic, Daniel Jiang, Aaditya Ramdas, Michael Jordan
Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics, PMLR 108:3806-3815, 2020.

Abstract

One important partition of algorithms for controlling the false discovery rate (FDR) in multiple testing is into offline and online algorithms. The first generally achieve significantly higher power of discovery, while the latter allow making decisions sequentially as well as adaptively formulating hypotheses based on past observations. Using existing methodology, it is unclear how one could trade off the benefits of these two broad families of algorithms, all the while preserving their formal FDR guarantees. To this end, we introduce Batch-BH and Batch-St-BH, algorithms for controlling the FDR when a possibly infinite sequence of batches of hypotheses is tested by repeated application of one of the most widely used offline algorithms, the Benjamini-Hochberg (BH) method or Storey’s improvement of the BH method. We show that our algorithms interpolate between existing online and offline methodology, thus trading off the best of both worlds.

Cite this Paper


BibTeX
@InProceedings{pmlr-v108-zrnic20a, title = {The Power of Batching in Multiple Hypothesis Testing}, author = {Zrnic, Tijana and Jiang, Daniel and Ramdas, Aaditya and Jordan, Michael}, booktitle = {Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics}, pages = {3806--3815}, year = {2020}, editor = {Chiappa, Silvia and Calandra, Roberto}, volume = {108}, series = {Proceedings of Machine Learning Research}, month = {26--28 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v108/zrnic20a/zrnic20a.pdf}, url = {https://proceedings.mlr.press/v108/zrnic20a.html}, abstract = {One important partition of algorithms for controlling the false discovery rate (FDR) in multiple testing is into offline and online algorithms. The first generally achieve significantly higher power of discovery, while the latter allow making decisions sequentially as well as adaptively formulating hypotheses based on past observations. Using existing methodology, it is unclear how one could trade off the benefits of these two broad families of algorithms, all the while preserving their formal FDR guarantees. To this end, we introduce Batch-BH and Batch-St-BH, algorithms for controlling the FDR when a possibly infinite sequence of batches of hypotheses is tested by repeated application of one of the most widely used offline algorithms, the Benjamini-Hochberg (BH) method or Storey’s improvement of the BH method. We show that our algorithms interpolate between existing online and offline methodology, thus trading off the best of both worlds.} }
Endnote
%0 Conference Paper %T The Power of Batching in Multiple Hypothesis Testing %A Tijana Zrnic %A Daniel Jiang %A Aaditya Ramdas %A Michael Jordan %B Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2020 %E Silvia Chiappa %E Roberto Calandra %F pmlr-v108-zrnic20a %I PMLR %P 3806--3815 %U https://proceedings.mlr.press/v108/zrnic20a.html %V 108 %X One important partition of algorithms for controlling the false discovery rate (FDR) in multiple testing is into offline and online algorithms. The first generally achieve significantly higher power of discovery, while the latter allow making decisions sequentially as well as adaptively formulating hypotheses based on past observations. Using existing methodology, it is unclear how one could trade off the benefits of these two broad families of algorithms, all the while preserving their formal FDR guarantees. To this end, we introduce Batch-BH and Batch-St-BH, algorithms for controlling the FDR when a possibly infinite sequence of batches of hypotheses is tested by repeated application of one of the most widely used offline algorithms, the Benjamini-Hochberg (BH) method or Storey’s improvement of the BH method. We show that our algorithms interpolate between existing online and offline methodology, thus trading off the best of both worlds.
APA
Zrnic, T., Jiang, D., Ramdas, A. & Jordan, M.. (2020). The Power of Batching in Multiple Hypothesis Testing. Proceedings of the Twenty Third International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 108:3806-3815 Available from https://proceedings.mlr.press/v108/zrnic20a.html.

Related Material