The application of machine learning (ML) in laser-induced breakdown spectroscopy (LIBS) has made LIBS a dominant technique for soil analysis. However, due to variation in the physical properties and/or the chemical composition of the sample materials, the spectral emission lines of elements can be highly variable, which cause different distributions of LIBS training and test spectra. For example, the laser shots targeted at different positions on the same sample material may introduce shifts in input distributions, which is usually the case in LIBS systems. Still, they are always ignored in the spectra analysis. When training and test distribution are different, applying an ML model learned from the training distribution to the test distribution violates the assumption for classical ML algorithms that the training and test data must come from the same distribution. Therefore, an ML model must be able to handle such domain shifts and adapt itself to the emission line distribution change...

Laser-induced breakdown spectroscopy (LIBS) has become an emerging analytical technique for soil analysis. The application of machine learning for quantitative and qualitative analysis has made LIBS more promising. However, the emission line distribution can be highly variable due to the soil samples' varied physical properties and/or chemical composition. It may cause spectra distribution change and make the training spectra not accurately reflect the test spectra distribution. Hence, the test performance is deteriorated by applying an ML model trained on samples from the training distribution to the test distribution. To handle the spectra distribution problem, we propose using pseudoshot learning with Siamese networks, a domain adaptation (DA) method to incorporate pseudolabeled samples based on similarity confidence into the parameter estimation procedure. Considering the domain transfer differences among classes, we categorize the classes as hard, normal, and easy to reflect the c...