Could be interesting to implement a class of "distribution-free" p-boxes, using just range and moment constraints. I.e., if we do not know the distribution of the uncertainty, but we have their range and moments.
Generally these p-boxes are constructed using classical probabilistic inequalities, like the Chebyshev and Cantelli inequalities.
See plots 1. and 2. and 3. and 4. of this paper: https://doi.org/10.1016/j.ijar.2022.04.001
Cdf bounds are can be found inside that paper, and the quantiles inside this software: https://github.com/AnderGray/ProbabilityBoundsAnalysis.jl
Could be interesting to implement a class of "distribution-free" p-boxes, using just range and moment constraints. I.e., if we do not know the distribution of the uncertainty, but we have their range and moments.
Generally these p-boxes are constructed using classical probabilistic inequalities, like the Chebyshev and Cantelli inequalities.
See plots 1. and 2. and 3. and 4. of this paper: https://doi.org/10.1016/j.ijar.2022.04.001
Cdf bounds are can be found inside that paper, and the quantiles inside this software: https://github.com/AnderGray/ProbabilityBoundsAnalysis.jl