A brief review of new data analysis methods of laser-induced breakdown spectroscopy: machine learning

A tutorial review on methods of agricultural product sample pretreatment and target analysis by laser-induced breakdown spectroscopy

“…These methods are described in more detail in the review. 13 These studies demonstrate that LIBS can easily be used to determine if harmful chemicals in products such as grains, fruits, and vegetables exceed the limit values. Furthermore, due to the heterogeneity of agricultural products, the sensitivity of detecting trace mineral elements and heavy metals at low concentrations in complex organic matter still needs to be improved.…”

“…In recent years, many researchers have reviewed and discussed the current status and prospects of LIBS applications in various elds, 7 such as food analysis, 8,9 soil analysis, 10,11 metal oxides 12 and spectral data analysis methods. 13 Several scholars reviewed LIBS and XRF in plant analysis. 14 Meanwhile, some reviews 15,16 were conducted in the eld of agriculture.…”

Atomic spectrometry update: review of advances in atomic spectrometry and related techniques

“…The authors considered that this review could shed light on further developments necessary for ANN-based LIBS chemometrics in the future. Zhang et al 136 also reviewed progresses on the application of machine learning algorithms to LIBS and evaluated the problems and challenges of their application.…”

Optimization of spodumene identification by statistical approach for laser-induced breakdown spectroscopy data of lithium pegmatite ores

“…A LIBS spectrum of a certain mineral can be considered as a fingerprint within a dataset: A unique composition of emission lines with characteristic line intensity values. This enables the use of chemometric data analysis for the characterization of the LIBS spectra and an extensive review is presented elsewhere (29). In the case of mineral and rock samples several LIBS studies are published with varying techniques, e.g., principal component analysis (PCA) (30)(31)(32)(33)(34)(35), partial least squares (PLS) (36), partial least squares discriminant analysis (PLS-DA) (30,34,(37)(38)(39), hybrid sparse partial least squares (SPLS) and leastsquares support vector machine (LS-SVM) model ( 40), K-means clustering (41,42), soft independent modelling of class analogy (SIMCA) (35,41), support vector machines (SVM) (39,43), singular value decomposition (SVD) (44), convolutional neural network with twodimensional input (2D CNN) (45), Spectral Angle Mapper (SAM) (46), and random forest (RF) (47).…”