Calibration Model Updating to Novel Sample and Measurement Conditions without Reference Values

Abstract

Updating a calibration model formed in original (primary) sample and spectral measurement conditions to predict analyte values in novel (secondary) conditions is an essential activity in analytical chemistry in order to avoid a complete recalibration. Established model updating methods require sample analyte reference values for a small set of secondary domain samples (labeled data) to be used in updating processes. Because obtaining reference values is time consuming and is the costly part of any calibration, methods are needed that do not require labeled secondary samples, thereby allowing on demand model updating. This paper compares model updating methods with and without labeled secondary samples. A hybrid model updating approach is also developed and evaluated. Unfortunately, a major impediment to adapting a model without secondary analyte reference values has been model selection. Because multiple tuning parameters are commonly involved in model updating methods, thousands of models are formed, making model selection complex. A recently developed framework is evaluated for automatic model selection of several two to three tuning parameter-based model updating methods without secondary analyte reference values (labels). The model selection method is based on model diversity and prediction similarity (MDPS) of the unlabeled samples to be predicted. The new secondary samples to be predicted can be used to form the updated models and again to select the final predicting models. Because models are formed and selected on demand to directly predict target samples, complicated cross-validation processes are not needed. Four near-infrared data sets covering 40 model updating situations are evaluated showing that MDPS can select reliable updated models outperforming or rivaling prediction errors from total recalibrations with secondary reference values.

Physicochemical Responsive Integrated Similarity Measure (PRISM) for a Comprehensive Quantitative Perspective of Sample Similarity Dynamically Assessed with NIR Spectra

“…The initial LV range is from 1 LV to sufficiently overdetermined, e.g., the 99% rule. Before model selection by MDPS, weight and LV values are dynamically resized per data set to remove model quality zones of convergence. , …”

“…Many model updating algorithms exist ranging from expensive requiring some target domain sample analyte reference values, 11,32 to cheap requiring no target sample reference values. 29,31,32 The model updating application assesses updating difficulty relative to model generalizability, an assessment problem that to date, lacks any potential solution. 28 The model updating methods local mean centering (LMC) 32 and null augmented regression eigenvalue (NARE) 29,32 are used to evaluate PRISM and the reader is referred to the respective references for details on the methods.…”

“…If a target sample is deemed a nonmember of the AD, recourse is available by calibration transfer using transfer learning methods. , These techniques adapt the calibration model to account for the novel target domain conditions, either by spectral preprocessing or direct model adaptation by model updating. , Model updating reorients the model in direction and magnitude to accurately predict the source and target samples. In terms of eq S5 in the SI, model updating attempts to alter the model b̂ such that the nonanalyte part of u matched to source and target sample matrix effect differences s goes to zero.…”

“…Before model selection by MDPS, weight and LV values are dynamically resized per data set to remove model quality zones of convergence. 29,30 Software. Algorithms were developed by the authors using MATLAB 2022a and the Parallel Computing Toolbox.…”

“…In terms of eq S5 in the SI, model updating attempts to alter the model b̂ such that the nonanalyte part of u matched to source and target sample matrix effect differences s goes to zero. Many model updating algorithms exist ranging from expensive requiring some target domain sample analyte reference values, , to cheap requiring no target sample reference values. ,, …”