PREDNASAJUCI / LECTURER : Július Koza (1) Astronomical Institute, Slovak Academy of Sciences, 059 60 Tatranská Lomnica NAZOV / TITLE : Bayesian Inversion of Solar Synthetic Off-Limb Spectra – Test Results ABSTRAKT / ABSTRACT : Solar physics employs a wide range of inversion methods and codes, primarily for interpreting spectropolarimetric observations of the photosphere and chromosphere on the solar disk. However, inversion techniques for off-limb spectra are limited, and no existing tool allows for the simultaneous inversion of key Balmer and Ca II lines. This talk presents test results from an inversion experiment where synthetic emission lines H-beta and Ca II 8542 Å were analyzed using a Bayesian probabilistic approach. Synthetic emission profiles of H-beta and Ca II 8542 Å were computed for a prominence modeled as a 1D vertical slab with finite thickness but infinite width and height. The radiative transfer problem in non-LTE conditions was solved using the Multi-Level Accelerated Lambda Iteration (MALI) method separately for hydrogen and calcium. The slab model is characterized by four free parameters: temperature, gas pressure, microturbulent velocity, and slab thickness. Bayesian inference was implemented using the Solar Bayesian Analysis Toolkit (SoBAT), with parameter space sampling performed via the Markov Chain Monte Carlo (MCMC) method using the MCMC_FIT.PRO routine from SoBAT. For an initial sample size of 10,000, the inversion successfully recovered the original parameter set in 7 out of 10 test cases. Increasing the sample size improved the success rate to 9 out of 10. Convergence to the correct solution typically required 1.5 to 2 days. However, the tests revealed an unexpected degeneracy between slab thickness and gas pressure, where significantly different values of these parameters produced nearly identical synthetic profiles of H-beta and Ca II 8542 Å. This correlation poses a serious challenge. Addressing this issue, along with optimizing the inversion process for efficiency, will be the focus of future work.