In this study,
a new approach to laser-induced breakdown spectroscopy
(LIBS) data modeling using multiway algorithms was investigated. Two
case studies, parallel factor analysis (PARAFAC) and unfolded-partial
least-squares with residual bilinearization (U-PLS/RBL) algorithms
were used in (1) the determination of Al, Cu, and Fe in samples of
reference material of printed circuit board (PCB) from electronic
waste and (2) the determination of Ca, K, and Mg in samples of a human
mineral supplement, where depth was used to obtain multidimensional
data in the first case and delay-time in the second. In addition,
univariate calibration was applied and compared with the multiway
approaches. In all cases, the calibration data set was prepared from
salts. PARAFAC showed satisfactory results in the first study, with
low prediction errors and good accuracy for most samples, and the
U-PLS/RBL algorithm presented the best performance for mineral supplement
samples.