For a long time, the Mythbusters have harboured a grudge about the ubiquitous use of and blind reliance on the RPD statistic as a tell-all index regarding the quality of any NIR calibration. This feeling is substantiated by the many reviews we, and many others, have performed for the NIR community and other journals over more than the last decade. There are apparently very few reservations regarding this institution: if only RPD>3… all is well with prediction performance!
EsbensenK.H.GeladiP., “Principles of proper validation: Use and abuse of resampling for validation”, J. Chemometr.24, 168–187 (2010). doi: http://dx.doi.org/10.1002/cem.1310
4.
EsbensenK.H.JuliusL.P., “Representative sampling, data quality, validation–a necessary trinity in chemometrics”, in Comprehensive Chemometrics, Ed by BrownS.TaulerR.WalczakB., Elsevier, pp. 1–20 (2009).
5.
MartensH.NæsT., Multivariate Calibration. Wiley, Chichester, UK (1989).
6.
EsbensenK.H.WagnerC., “Theory of sampling (TOS) vs. measurement uncertainty (MU)–a call for integration”, Trends Anal. Chem. (TrAC)57, 93 (2014). doi: http://dx.doi.org/10.1016/j.trac.2014.02.007
7.
EsbensenK.H.PaolettiC.MinkkinenP., “Representative sampling of large kernel lots–I. Theory of Sampling and variographic analysis”, Trends Anal. Chem. (TrAC)32, 154 (2012). doi: http://dx.doi.org/10.1016/j.trac.2011.09.008
8.
SørensenM.LarsenA.EsbensenK.H., “Visualisation of sampling error effects in near infrared analysis–comparison between Petri dish, roll bottle and spiral sampler”, NIR news25(1), 11 (2014). doi: http://dx.doi.org/10.1255/nirn.1414
