Description Usage Format Source See Also
The simulated data set simul2
considers a situation with clustered
observations and four binary covariates in both sub-models of the Pogit
model, i.e.
X
= W
. The respective design matrix is
built by computing all 2^4 possible 0/1 combinations and one observation
is generated for each covariate pattern. C=50 clusters are built containing
one unit with each of the resulting 16 covariate patterns, i.e. a total of
I=800 units. The regression effects are set to beta = {0.75,0.1,0.1,0,0}
in the Poisson and to alpha = {2.2,-0.3,0,-0.3,0}
in the logit model.
Random intercepts in both sub-models are simulated from a normal distribution
with standard deviations θ_β=0.1
and
θ_α=0.3
. Additionally to the main study sample,
validation data are available for each covariate pattern and cluster.
For details concerning the simulation setup, see Dvorzak and Wagner (2016).
1 |
A data frame with 800 rows and the following 10 variables:
y
number of observed counts for each covariate pattern in each cluster
E
total exposure times for each unit
cID
cluster ID for each unit
X.0
intercept
X.1
, X.2
, X.3
, X.4
binary covariates
v
number of reported cases for each covariate pattern in each cluster in the validation sample
m
number of true cases subject to the fallible reporting process (sample size of validation data)
Dvorzak, M. and Wagner, H. (2016). Sparse Bayesian modelling of underreported count data. Statistical Modelling, 16(1), 24 - 46, http://dx.doi.org/10.1177/1471082x15588398.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.