1 January 2019 in TrAC Trends in Analytical Chemistry
25 February 2026 in Journal of Analytical Atomic Spectrometry
With the increasing demand for marine resources, the existing marine exploration technologies face multiple challenges.
28 July 2026 in SSRN Electronic Journal
The miniaturization and integration of laser-induced breakdown spectroscopy (LIBS) technology is one of its important development trends, and also the only way for it to be practical application from the laboratory. Aiming at the problems of limited computing power and relatively low model computing efficiency in the qualitative analysis of compact LIBS technology, a feature selection method based on the cumulative correlation coefficient (CCC) was proposed. It was verified by the classification based on linear discriminant analysis (LDA) algorithm for 12 kinds of igneous rock samples. Using this method, the classification accuracy (CA) and relative standard deviation (RSD) of the samples reached 94.8% and 12.4%, respectively, while only 91 analytical lines were required in its qualitative analysis model. Compared with traditional signal-to-noise ratio (SNR) (105 Lines) or signal-to-background ratio (SBR) (210 lines) methods, the feature dimension of the model was effectively reduced by 13.3% and 56.7%, respectively. The time complexity and space complexity have also been significantly reduced. It is of great significance for improving the analysis efficiency of the compact LIBS system. It provided a way that can be standardized and produced deterministic results for the feature selection process of LIBS, and ensured the reliability of the analysis model and result.