calc.p computes p (proportion) parameter from a
and lambda coefficients in a broken stick model.
build.p.lambda parses the x vector, usually returned by
the coef method, where \(x =
(p_0,\dots,p_n,\lambda_1,\dots,\lambda_{n+1})\),
and build a named list with p and lambda elements to use
in fitting functions.
logit and unLogit are helpful for reparameterizing the
negative maximum likelihood function, if using Langton et al. (1995).
calc.p(coefs)build.p.lambda(x)
logit(p)
unLogit(logit)
numeric vector with proportion parameters implied by
coefs.
named (p, lambda) list with parsed coefficients.
unLogit and logit return a numeric vector with
the (un)transformed arguments.
numeric matrix [2,N] of coefficients (a and
lambda) in rows for each process of the model in columns.
Columns are assumed to be in decreasing order with respect to
lambda
numeric vector of coefficients
numeric vector of proportions (0-1) to transform to the logit scale.
numeric scalar: logit value to transform back to original scale.
Sebastian P. Luque spluque@gmail.com