Assessing Confidence in Predictions Using Veracity and Utility – A Case Study on the Prediction of Mammalian Metabolism by Meteor Nexus. Issue 5 (24th April 2015)

• Record Type: Journal Article
• Authors: Judson, Philip N.; Long, Anthony; Murray, Ernest; Patel, Mukesh
• Is Part Of: Molecular informatics. Volume 34:Issue 5(2015:May)
• Page Start: 284
• Page End: 291
• Publication Date: 2015-04-24
• Subjects: Cheminformatics -- Periodicals; QSAR (Biochemistry) -- Periodicals; Structure-activity relationships (Biochemistry) -- Periodicals; Drugs -- Structure-activity relationships -- Periodicals; 615.19
• DOI: 10.1002/minf.201400184
• Languages: English
• ISSNs: 1868-1743

Abstract

<abstract abstract-type="main" xml:lang="en"> <title>Abstract</title> <p>A previous paper<xref ref-type="link" rid="bib1">1</xref> described new metrics, veracity and utility, for assessing the performance of toxicity prediction systems that report confidence in their predictions. Assessing the performance of systems that predict mammalian metabolism is complicated by the absence of comprehensive sets of negative observations and predictions. This paper presents an approach to assessing the performance of such systems using veracity and utility.</p> </abstract>