Skip to content

Latest commit

 

History

History
5 lines (5 loc) · 850 Bytes

Simulation_Design.md

File metadata and controls

5 lines (5 loc) · 850 Bytes

How to prove that a method based on the A probabilistic distribution shows performance comparable to or better compared to another method based on the B probabilistic distribution?

  • A/B: Normal/Binomial/Poisson/Gamma/Negative binomial/Log-normal/Hypergeometric/Laplace/Pareto/Dirichlet/Leptokurtic...
  • Design simulations where the values of a variable of interest are generated using the B probabilistic distribution as another analytical model assumes.
  • These simulations will be the benchmark for the method based on the B probabilistic distribution.
  • If an approach based on the A probabilistic distribution can demonstrate performance equal to or surpassing a method based on the B probabilistic distribution in these simulations, it strongly suggests its competitiveness across the data generated from the B probabilistic distribution.