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


Current browse context:


Change to browse by:

References & Citations


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

Statistics > Applications

Title: Inference for Hit Enrichment Curves, with Applications to Drug Discovery

Abstract: In virtual screening for drug discovery, hit enrichment curves are widely used to assess the performance of ranking algorithms with regard to their ability to identify early enrichment. Unfortunately, researchers almost never consider the uncertainty associated with estimating such curves before declaring differences between performance of competing algorithms. Appropriate inference is complicated by two sources of correlation that are often overlooked: correlation across different testing fractions within a single algorithm, and correlation between competing algorithms. Additionally, researchers are often interested in making comparisons along the entire curve, not only at a few testing fractions. We develop inferential procedures to address both the needs of those interested in a few testing fractions, as well as those interested in the entire curve. For the former, four hypothesis testing and (pointwise) confidence intervals are investigated, and a newly developed EmProc approach is found to be most effective. For inference along entire curves, EmProc-based confidence bands are recommended for simultaneous coverage and minimal width. Our inferential procedures trivially extend to enrichment factors, as well.
Comments: 34 pages, 6 figures, 2 tables. Original version was a chapter in Jeremy Ash's dissertation. Current version is the paper submitted to Journal of American Statistical Association: Applications and Case Studies
Subjects: Applications (stat.AP); Information Retrieval (cs.IR); Machine Learning (cs.LG); Quantitative Methods (q-bio.QM); Methodology (stat.ME)
Cite as: arXiv:1912.09526 [stat.AP]
  (or arXiv:1912.09526v3 [stat.AP] for this version)

Submission history

From: Jeremy Ash [view email]
[v1] Thu, 19 Dec 2019 20:08:03 GMT (6126kb,D)
[v2] Wed, 22 Sep 2021 14:07:33 GMT (583kb,D)
[v3] Thu, 23 Sep 2021 16:30:43 GMT (584kb,D)

Link back to: arXiv, form interface, contact.