Altered, Inflated and Truncated Values in GAIT Regression
Return the altered, inflated and truncated values in a GAIT regression object, else test whether the model is altered, inflated or truncated
altered(object, ...) inflated(object, ...) truncated(object, ...) is.altered(object, ...) is.inflated(object, ...) is.truncated(object, ...)
object |
an object of class |
... |
any additional arguments, to future-proof this function. |
Yee and Ma (2020) propose GAIT regression where values
from three disjoint sets are referred to as special.
These extractor functions return one set each;
they are the alter, inflate, truncate
(and sometimes max.support)
arguments from the family function.
Returns one type of ‘special’ sets associated with GAIT regression.
This is a vector, else a list for truncation.
All three sets are returned by specialsvglm.
Some of these functions are subject to change.
Only family functions beginning with "gait" will
work with these functions, hence
zipoisson fits will return FALSE or empty
values.
Yee, T. W. and Ma, C. (2020). Generally–altered, –inflated and –truncated regression, with application to heaped and seeped counts. In preparation.
abdata <- data.frame(y = 0:7, w = c(182, 41, 12, 2, 2, 0, 0, 1))
fit1 <- vglm(y ~ 1, gaitpoisson(alt.mix = 0),
data = abdata, weight = w, subset = w > 0)
specials(fit1) # All three sets
altered(fit1) # Subject to change
inflated(fit1) # Subject to change
truncated(fit1) # Subject to change
is.altered(fit1)
is.inflated(fit1)
is.truncated(fit1)Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.