profile.simode {simode} | R Documentation |
Calculate likelihood profiles for the model parameters
Description
Calculate likelihood profiles for the model parameters
Usage
## S3 method for class 'simode'
profile(
fitted,
which = NULL,
optim_type = c("nls", "both"),
step_size,
max_steps = 100,
alpha = 0.05,
skip_err = T,
trace = 0,
save_to_log = F,
...
)
Arguments
fitted |
|
which |
Which parameters to estimate the profile for. |
optim_type |
Whether to calculate the profiles based on maximum-likelihood optimization only ('nls') or based on integral-matching followed by maximum-likelihood optimization ('both'). |
step_size |
Step size for profiling (one value for all parameters or
a value for each parameter in |
max_steps |
Maximum number of steps to take in each direction. |
alpha |
Maximum (two-sided) likelihood ratio test confidence level to find. |
skip_err |
Whether on not to stop the calculation if encountering a problem with one point in the profile. |
trace |
Report level (0-4), with higher values producing more tracing information. |
save_to_log |
Whether to redirect output to log file. The log file will be saved to tempdir(). |
... |
Additional argument(s) for methods. |
Details
If the call to simode
, which returned the fitted object given to
this method, included a user-defined likelihood function (with the calc_nll
argument),
then the likelihood profiles will be calculated using this function. Otherwise,
the profiles will be calculated using a likelihood based on a Gaussian distribuion
with fixed sigma, where sigma will be estimated in the background together with
the rest of the model parameters.
Value
The likelihood profiles.