Fig. 1
From: Mixture density networks for the indirect estimation of reference intervals
Simulation example. Left: Example draw of 5000 points from the data-generating process. Samples from the “main” component are colored black and backgrounded with an area depicting the density of their parent distribution. Grey points are sampled from the “pathologic” component. Right: True parameters of the data-generating process. Regarding the mixture component weights, the dependent and independent scenarios are depicted by solid and dashed lines, respectively