1：Experimental Design and Method Selection:

The study utilizes a feature selection method called recursive feature elimination based on ridge regression (Ridge-RFE) for the original spectral data to make full use of the valid information of spectra. The absolute value of the ridge regression coefficient is used as a criterion to screen spectral characteristic.

2：Sample Selection and Data Sources:

Fifty-one standard aluminum alloy samples were used for the analysis. The spectral signal is a 1× 4094 dimensional vector collected using the AvaSpec-Mini spectrometer.

3：List of Experimental Equipment and Materials:

The experimental device is a portable LIBS analyzer developed by the Shenyang Institute of Automation, Chinese Academy of Sciences. The excitation light source is YLP-C-20 fiber laser from China Jiepu.

4：Experimental Procedures and Operational Workflow:

Each sample is measured at 10 different positions, and the average spectrum of 22 spectra is taken as the spectrum obtained in one measurement. The final spectral data is 51

5：Data Analysis Methods:

40 The determination coefficients of cross-validation, the root mean square error of cross-validation (RMSECV) and the root mean square error of prediction (RMSEP) were used to assess the performance of the calibration and prediction qualities of the model.