- Maintenance for new S3 method lookup in R 4.0.0.
`NextMethod("confint")`

in `confint.fitC()`

wouldn’t find a method if `confint.fitC()`

was called
directly from `summary.fitC()`

.

This is the first version published on CRAN. It requires **surveillance**
>= 1.14.0. The main change is in the documentation, which became more
comprehensive, especially for the included datasets.

Other changes include:

Changed the default `agegroups`

parameter in
`noroBE()`

(and `grouping`

parameter in
`contactmatrix()`

) such that it corresponds to the six age
groups used by Meyer and Held (2017).

The `aggregateC()`

function is now exported.

This version has been used for the publication

Held L, Meyer S and Bracher J (2017). “Probabilistic forecasting in
infectious disease epidemiology: The 13th Armitage lecture.”
*Statistics in Medicine*, doi: 10.1002/sim.7363.

## New features

An additional year of norovirus surveillance counts has been
added, such that `noroBE()`

can now extract data from the
time range 2011-w01 to 2016-w30.

`noroBE(by = "none")`

extracts the overall univariate
time series.

`update()`

method for `"fitC"`

objects,
which enables modification of the `"hhh4"`

model with
automatic refitting of the power parameter of the contact
matrix.

`plotHHH4_season_groups()`

gained an option to disable
confidence intervals.

## Bug fixes

- A typical
`drop=FALSE`

bug prevented
`aggregateC()`

from aggregating a contact matrix to only two
groups.

This is the original version published as supplementary data of

Meyer S and Held L (2017). “Incorporating social contact data in
spatio-temporal models for infectious disease spread.”
*Biostatistics*, **18**(2), pp. 338-351, doi:
10.1093/biostatistics/kxw051.