BIC {qpcR} | R Documentation |
Calculates the Bayesian Information Criterion for objects of class drc
, lm
, glm
, nls
or any other models from which logLik
, coef
and residuals
can be extracted.
BIC(object)
object |
a fitted model. |
BIC = - 2 * logLik(object) + npar * log(nobs)
with npar = number of parameters, nobs = number of observations.
The BIC value.
Andrej-Nikolai Spiess
Schwartz GE (1978).
Estimating the dimension of a model.
Annals of Statistics, 6: 461-464.
m <- pcrfit(reps, 1, 2, l5) BIC(m)