Interval estimation of small tail probabilities – applications in food safety

Benjamin Kedem, Lemeng Pan, Wen Zhou, Carlos A. Coelho

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Often in food safety and bio-surveillance it is desirable to estimate the probability that a contaminant or a function thereof exceeds an unsafe high threshold. The probability or chance in question is very small. To estimate such a probability, we need information about large values. In many cases, the data do not contain information about exceedingly large contamination levels, which ostensibly renders the problem insolvable. A solution is suggested whereby more information about small tail probabilities are obtained by combining the real data with computer-generated data repeatedly. This method provides short yet reliable interval estimates based on moderately large samples. An illustration is provided in terms of lead exposure data. Copyright © 2016 John Wiley & Sons, Ltd. Copyright © 2016 John Wiley & Sons, Ltd.
Original languageEnglish
Pages (from-to)3229-3240
JournalStatistics in Medicine
Volume35
Issue number18
DOIs
Publication statusPublished - 2016

Keywords

  • coverage
  • density ratio model
  • food safety
  • NHANES
  • out of sample fusion
  • semiparametric

Fingerprint Dive into the research topics of 'Interval estimation of small tail probabilities – applications in food safety'. Together they form a unique fingerprint.

Cite this