The Matlab file LHSPRCC.m is the main code file which calls the function DrawSamples.m to perform the Latin hypercube sampling step, any user-specified model functions for completing the Monte-Carlo Simulations, and either UnariedPRCC.m or VariedPRCC.m to compute partial rank correlation coefficients (at a single time/location index or at all times/locations). This repository contains code to conduct LHS+PRCC analysis in either matlab or python, depending on user preference. ![]() A brief illustration of utility of this method as applied to the proliferation-invasion-recruitment model will be on BioRxiv (as part of the mathematical oncology channel) in the near future. The LHS method for parameter sampling in Monte Carlo studies was first developed by McKay, Beckman, and Conover, 1979 and was applied in conjunction with partial rank correlation coefficients for use in biomathematical models in Blower and Dowlatabadi 1994. This can be useful in developing the model to understand how it behaves in various parameter regimes, as well as to understand better how uncertainty in your parameter estimates may impact the results given by the model.Īn overview of the procedure is provided as a pdf slide deck. ![]() LHS + PRCC is a useful method for investigating the sensitivity of a mathematical model to it's parameters. Model-sensitivity-analysis Latin hypercube sampling and partial rank correlation coefficients for analyzing model parameter sensitivity.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |