Second- and third-order multivariate calibration: data, algorithms and applications

Extent-based kinetic identification using spectroscopic measurements and multivariate calibration

“…In absorption spectroscopy, calibration models based on Principal Component Regression (PCR) [15] or Partial Least Squares (PLS) [16] are used to predict the concentrations of absorbing species from an absorbance spectrum measured in a sample [17,18]. Calibration involves two steps.…”

Chemometric strategies for enhancing the chromatographic methodologies with second-order data analysis of compounds when peaks are overlapped

“…Combining the three-way array of N-PLS with the RBL procedure in order to exploit the second-order advantage N-PLS/RBL [10] From an analytical point of view, the main object of using second-order data is to assure exploiting the second-order advantage. For this purpose, two alternative ways are possible [6]: (a) data for the test sample has an influence on the regression coefficients leading to prediction (PARAFAC, PARAFAC2, GRAM, DTLD, APTLD, SWATLD and MCR-ALS) or (b) calibration is first performed using only training data, with the test sample leading to sample-specific regression coefficients in a subsequent step (BLLS, U-PLS/RBL, NPLS/RBL and ANN/RBL).…”

“…In addition, complete reviews were presented with a wide range of applications to second-order data (including chromatography) [6][7][8]. Thus, in this review we will revise the most recent and important applications, especially in those cases in which complex samples were analyzed and overlapping peaks had to be solved because of incomplete selectivity due to insufficient separation among analytes or the presence of unexpected components in the sample.…”

“…Chemometrics is especially useful in chromatography when second-order data are recorded, for example, using a diode-array detector (DAD), a fast-scanning luminescence or a mass spectrometry based (MS) detector during the chromatographic time evolution. An interesting intrinsic property that secondorder data may show (if they are modeled with convenient second-order algorithms) is the so-called "second-order advantage" [5], which in principle permits analyte quantitation in samples containing unexpected components, i.e., compounds not included in the calibration set [6]. This fact allows one to build a predictive model with a limited number of standards, yet quantitating the analyte in the presence of potential interferents [6].…”

“…The use of second-order multivariate algorithms has been shown to play a critical role in several analytical fields, as can be gathered from a literature survey in relevant analytical, chemometrics and applied journals [6,7]. Specifically, an important number of reports have been presented ocusing on the resolution of really complex samples using liquid chromatography and exploiting the mentioned second-order advantage [8][9][10][11][12].…”