MOST: Markov Ordinal State Transition model
Alternative to GEE
MOST: Markov Ordinal State Transition model
Alternative to GEE
Common choices of F,
includes list of respective link functions. page 362.
Dirichlet Priors With Concentration Parameter 0.148 for Intercepts
Prior on the Intercepts: depend on number of levels; automatically estimated by the function.
Odds ratio capable of being constant Ex: risk factor doubles odds of disease
Great visualization OR vs Risk difference
at least one of the training and test samples is not huge
sample size should be not smaller than 20000 before applying split-based strategy
rexVar function
calculate relative explained variance; see also proportion of chisq for each predictor.
Strong internal validation using 100 repeats of 10-fold cross-validation or several hundred bootstrap resamples, repeating all analysis steps involving Y afresh at each re-sample and the arbitrariness of selected “important variables” is reported (if variable selection is used)
also appropriate for small sample sizes
Predictions plot from Rousseeuw (2026)
Does not harm predictions on model construction sample
about collinearity - as long as collinearity changes across new samples, this holds also for new samples
Breiman’s nonnegative garrote (Xiong (2010))
scale-free
understand influence of knot placement
cyclical trends
example of modeling cyclical trends