Internal functions called by [transreg()], depending on choice between exponential and isotonic calibration.
.exp.multiple(
y,
X,
prior,
family,
switch = FALSE,
select = TRUE,
track = FALSE
).iso.multiple(
y,
X,
prior,
family,
switch = FALSE,
select = TRUE,
track = FALSE
)
.iso.fast.single(y, X, prior, family)
.iso.slow.single(y, X, prior, family)
target: vector of length \(n\) (see family)
features: matrix with \(n\) rows (samples) and \(p\) columns (features)
prior coefficients: matrix with \(p\) rows (features) and \(k\) columns (sources of co-data)
character "gaussian" (\(y\): real numbers), "binomial" (\(y\): 0s and 1s), or "poisson" (\(y\): non-negative integers);
choose between positive and negative weights for each source: logical
select from sources: logical
show intermediate output (messages and plots): logical
.exp.multiple(): called by `transreg` if `scale="exp"`
.iso.multiple(): called by `transreg` if `scale="iso"`
.iso.fast.single(): called by `transreg` if `scale="iso"` (via `.iso.multiple`)
.iso.slow.single(): replaced by `.iso.fast.single`
Use [transreg()] for model fitting.