The long form documentation points to the use of RKF78 as the numerical integrator, yet DAISIE v4.1.1 uses lsodes as default. Should this be changed to match DAISIE?
The details of the maximum likelihood set up are not important, but a brief explainer: all model parameters are optimisied with the starting position in parameter space for optimisation, equal to the values used to simulate the data; the Runge-Kutta Fehlberg method to numerically solve the likelihood equations; and simplex is the optimisation algorithm to maximise the likelihood.
RKF78 was used for all analyses and is explicitly stated as the methode argument throughout. Other people can change these settings if they wish, but let's leave as is for the documentation for consistency.
The long form documentation points to the use of RKF78 as the numerical integrator, yet DAISIE v4.1.1 uses lsodes as default. Should this be changed to match DAISIE?
See here: https://joshwlambert.github.io/DAISIEmainland/inference-performance.html#maximum-likelihood-inference
emphasis mine.