Bayesian classification maps the probable occurrence of endmember facies from inverted seismic. The prior distributions is typically well logs, from which distinct lithology and fluid properties have been defined. Endmember separation from upscaled well logs is then mapped to inverted seismic property of choice. An example workflow is shown below

Stochastic workflows for uncertainty analysis involve systematically modelling and quantifying the impact of randomness and uncertainty in input parameters on model outputs, often using probabilistic methods and Markov Chain Monte Carlo techniques. These workflows include steps like identifying sources of uncertainty, propagating them through the model to understand their influence, and performing sensitivity analysis to determine their relative importance on the final prediction.