Stark spectral line broadening modeling by machine learning algorithms

New perspectives in the analysis of Stark width regularities and systematic trends

“…For more complex spectra, these estimates should be improved, or some other methods are welcome to be used. Furthermore, although RF model shows very strong potential to be applied on RST analysis in future, it is tested only in the sample of Stark broadening parameters related to simple spectra described here, and in the case of Li I spectral lines (Tapalaga et al, 2022), so it should also be confirmed in a greater sample to make us sure that this method can be applied generally in prediction of new Stark widths despite of complexity of a spectrum we investigate.For the application of these methods to study regularities and systematic trends among the Stark broadening parameters of lines in more complex spectra, additional investigations are needed, and development of both of these method are neccessary. Created database used in this and previous study is published online and it is available for use.…”

“…As machine learning represents a very popular tool for different types of problems encountered in science, here it was applied on the study of regularities of Stark broadening. Machine learning model based on Random Forest algorithm was developed and described in detail in reference (Tapalaga et al, 2022), so here it would be briefly described for the sake of completeness. Before developing the model, we needed to develop and create a database for training and testing of the future models.…”