An S3 method for extracting the regression coefficients of the
bsplines and btensor terms. By Default this uses
stats::coef to extract all the regression coefficients. A specific
method for lmerMod objects has been provided. If you are using a
regression method which stats::coef will not return the regression
coefficients, you'll need to define an S3 method for stats::coef to do
so.
Usage
coef_vcov(fit, theta_idx)
Arguments
- fit
a regression model fit
- theta_idx
numeric index for the theta related coefficients
Value
A list with four elements
- theta
theta regression coefficients
- coef
all regression coefficients
- vcov_theta
subsection of variance-covariance matrix pertaining to the theta values
- vcov
full variance-covariance matrix
Details
These functions are called in the cp and
cn calls.
Examples
cp0 <- cp(log10(pdg) ~ bsplines(day, df = 6, bknots = c(-1, 1)) + age + ttm, data = spdg)
cv <- cpr:::coef_vcov(cp0$fit)
summary(cv)
#> Length Class Mode
#> theta 8 -none- numeric
#> coef 8 -none- numeric
#> vcov_theta 64 -none- numeric
#> vcov 64 -none- numeric