interpret.logbin.smooth: Interpret a logbin.smooth Formula
In logbin: Relative Risk Regression Using the Log-Binomial Model

Description Usage Arguments Value Author(s) See Also Examples

View source: R/interpret.logbin.smooth.r

This is an internal function of package logbin. It is a service routine for logbin.smooth which interprets the smooth parts of the model formula and returns modified formulas to be used in the fitting functions.

Not normally called directly.

1	interpret.logbin.smooth(formula)

formula

A formula as supplied to logbin.smooth, which includes at least one B or Iso term.

A list with components:

`full.formula`	a `formula` object which is the same as the `formula` supplied, but with additional arguments removed from the smooth terms. E.g. `B(x, knot.range = 0:2)` would appear as `B(x)` in this formula.
`fake.formula`	a `formula` object which is the same as the `formula` supplied, but with smooth terms replaced by their covariates alone. E.g. `B(x, knot.range = 0:2)` would appear as `x` in this formula. Used to construct the model matrix.
`smooth.spec`	a named list containing the results of `eval`uating the smooth terms. See `B` and `Iso` for details.
`smooth.ind`	a vector containing the indices of the smooth components in the `formula`.
`terms`	the result of running `terms.formula(formula, specials = c("B", "Iso"))`.

Mark W. Donoghoe markdonoghoe@gmail.com

logbin.smooth

# Specify a smooth model with knot.range
res <- interpret.logbin.smooth(y ~ B(x, knot.range = 0:2) + x2)
# The knot.range is removed from the full.formula...
print(res$full.formula)
# ...but is stored in the $smooth.spec component of the result:
print(res$smooth.spec$x$knot.range)