Premium
Combining univariate calibration information through a mixed‐effects model
Author(s) -
Liao Jason J. Z.
Publication year - 2003
Publication title -
journal of chemometrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.47
H-Index - 92
eISSN - 1099-128X
pISSN - 0886-9383
DOI - 10.1002/cem.769
Subject(s) - univariate , chemometrics , calibration , multivariate statistics , statistics , mean squared error , set (abstract data type) , computer science , data mining , mixed model , mathematics , multivariate analysis , econometrics , machine learning , programming language
It is common practice to calibrate a common value by combining information from different sources such as days, people, instruments and laboratories. Under each individual source a univariate calibration can be used to calibrate the unknown. Then the common unknown can be estimated by combining the estimates from each source as a weighted mean (Johnson DJ, Krishnamoorthy K. J. Am. Statist. Assoc . 1996; 91 : 1707–1715) or through a multivariate calibration setting by combining information first and then estimating the common value (Liao JJZ. J. Chemometrics 2001; 15 : 789–794). In this paper a mixed‐effects model approach is proposed to combine good characteristics from both approaches. Simulations show that the mixed‐effects model has better bias and mean squared error (MSE) performance than the univariate and multivariate approaches. A real data set is used to demonstrate the good characteristics of the mixed‐effects model approach. Copyright © 2003 John Wiley & Sons, Ltd.