Abstract
The interpretation of the results of proficiency tests by the use of mixture models is described. The data are interpreted as a sample from a mixture of several normal populations. The calculation of the statistics (the means, variances and proportions of each component) is accomplished by means of the ‘EM’ algorithm. The method has several advantages over those previously advanced, principally that the algorithm is fast and easy to execute. Examples from proficiency testing are discussed.
Similar content being viewed by others
References
Lowthian PJ, Thompson M (2002) Analyst 127:1359–1364.
Cofino WP, van Stokkum IHM, Wells DE, Ariese F, Wegener JWM, Peerboom RAL (2000) Chemomet Intell Lab Syst 53:37–55
Thompson M, Wood R (1993) Pure Appl Chem 65:2123–2144. (Revision expected 2005)
Thompson M, Lowthian PJ (1997) J AOAC Internat 80:676–679
Fearn T (2004) Accred Qual Assur 9:441–444
Dempster AP, Laird NM, Rubin DB (1977) J Roy Stat Soc Series B 39: 1–38
Aitkin M, Wilson GT (1980) Technometrics 22:325–331
Thompson M, Accred Qual Assur (in press)
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Thompson, M. Using mixture models for bump-hunting in the results of proficiency tests. Accred Qual Assur 10, 501–505 (2006). https://doi.org/10.1007/s00769-005-0053-0
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00769-005-0053-0