Fig. 6

Probability of convergence (\({P}_{\text{fit\;valid}}\)) and probability of ‘correct’ estimation (\({P}_{\text{correct}}\)) (in percentage) obtained by simulations (black continuous lines) using the MCLM method, and using the PSO algorithm with the parameter configuration (dash-dotted lines) “Param. 1” used by Wu et al. (2015) and (dashed lines) “Param. 2”. Simulations were done at different signal fluctuation percentages (\(\delta (\%)\)) fitting different combinations of known a priori plasma parameters from the plasma line: without a priori information, given \({N}_{e}\), given \({N}_{e}\) and \({T}_{e}/{T}_{i}\), and given \({N}_{e}\) and \({T}_{e}\) parameters