lrpCI {longRPart} | R Documentation |
Using bootstrapping to calculate confidence intervals for the first split of a longitudinal regression tree, as outlined in the referenced paper. The details are outlined in the linked pdf.
lrpCI(model,B,alpha)
model |
model produced by longRPart |
B |
Number of bootstrap samples to run |
alpha |
The desired confidence level |
To calculate the confidence interval, B bootstrap samples of the patients are taken from the model data set, and a longitudinal regression tree is built, resulting B initial splitting values (one from each tree). The 100(1-a)% confidence interval is calculated using the quantiles of the B splitting values.
a vector of split values is returned, from which the confidence intervals can be obtained
Mohamed Abdolell <mo@dal.ca> and Sam Stewart
Abdolell et al. Binary partitioning for continuous longitudinal data: categorizing a prognostic variable. Statistics in Medicine (2002) vol. 21 pp. 3395-3409