The Bayesian methodology described in this paper has the inherentcapability of choosing, from calibration-type curves, candidateswhich are plausible with respect to measured data, expertknowledge and theoretical models (including the nature of themeasurement errors). The basic steps of Bayesian calibration arereviewed and possible applications of the results are described inthis paper. A calibration related to head-space gas chromatographicdata is used as an example of the proposed method. Thelinear calibration case has been treated with a log-normaldistributed measurement error. Such a treatment of noise stresses theimportance of modelling the random constituents of any problem.
展开▼
