Coefficient, covariance, and summary accessors for fits that store
independent per-arm coefficient blocks (cover_fit, occu_categorical_fit)
rather than a flat posterior. The arms are fit independently, so the joint
covariance is block-diagonal. When a full posterior is present (the NUTS
cover path) the inference methods defer to the draw-based tulpa_fit
methods.
Usage
# S3 method for class 'tobs_multiarm_fit'
nobs(object, ...)
# S3 method for class 'tobs_multiarm_fit'
coef(object, arm = NULL, ...)
# S3 method for class 'tobs_multiarm_fit'
vcov(object, ...)
# S3 method for class 'tobs_multiarm_fit'
confint(object, parm, level = 0.95, ...)
# S3 method for class 'tobs_multiarm_fit'
logLik(object, ...)
# S3 method for class 'tobs_multiarm_fit'
glance(x, ...)
# S3 method for class 'tobs_multiarm_fit'
tidy(x, conf.level = 0.95, ...)
# S3 method for class 'tobs_multiarm_fit'
summary(object, level = 0.95, ...)
# S3 method for class 'tobs_multiarm_fit'
plot(x, ...)
# S3 method for class 'tobs_multiarm_fit'
fitted(object, ...)
# S3 method for class 'tobs_multiarm_fit'
residuals(object, ...)Arguments
- object, x
A
tobs_multiarm_fit(acover_fitoroccu_categorical_fit).- ...
Ignored, or forwarded to
NextMethod()on the posterior path.- arm
Optional arm name (
"presence","positive"or"class") forcoef(); seecoef.tobs_fit().- parm
Ignored (present for
confint()generic compatibility).- level
Confidence level for
confint()andsummary()(default 0.95).- conf.level
Interval level for
tidy()(default 0.95).
Value
The layout every tobs_fit returns: nobs() an integer; coef() a
named vector (presence_(Intercept), positive_x, class_2:x, ...);
vcov() the block-diagonal covariance and confint() a two-column matrix,
both named as coef(); logLik() a logLik object; tidy() the
arm / term / estimate / std.error / conf.low / conf.high table;
summary() the estimate / std.error / interval data frame;
glance() the one-row column set of glance.tobs_fit(); plot() one
Gaussian-density panel per coefficient (up to 4), invisibly.