How to allow weighting of observations?
To weight your observations, say when using empirical logit, you must
provide a vector of variances, for each observation, which represents
the uncertainty of measurement. You can do this, provided the
family = "gaussian", add the argument, scale = vec, to fit_hts,
where vec is a vector of variances for each observation.
But I don't provide any worked examples of how to do this. I think the scale term belongs in the ... list at the end of the fit_hts function.
e.g.,
m <- fit_hts(
formula = pr ~ avg_lower_age + hts(who_subregion, country),
.data = malaria_africa_ts,
family = "gaussian",
special_index = month_num,
list(scale = vec)
)
Documentation currently says:
But I don't provide any worked examples of how to do this. I think the
scale
term belongs in the...
list at the end of thefit_hts
function.e.g.,