The Handbook of Infectious Disease Data Analysis (Chapman & Hall/CRC, 2019) contains a chapter on “Forecasting Based on Surveillance Data” (preprint) by Leonhard Held and Sebastian Meyer. The R package HIDDA.forecasting provides the data and code to reproduce results from the two applications presented in that chapter:
Univariate forecasting of Swiss ILI counts using
forecast::auto.arima() –> vignette("CHILI_arima")
glarma::glarma() –> vignette("CHILI_glarma")
surveillance::hhh4() –> vignette("CHILI_hhh4")
prophet::prophet() –> vignette("CHILI_prophet")
experimental kcde::kcde() –> vignette("CHILI_kcde")
and MASS::fitdistr() for naive reference forecasts –> vignette("CHILI_naive")
Age-stratified analysis of norovirus counts in Berlin using “hhh4” –> vignette("BNV")
To install the released HIDDA.forecasting package (with pre-built vignettes) together with other packages required to run the analyses:
# install.packages("remotes")
remotes::install_url("https://github.com/HIDDA/forecasting/releases/download/v1.1.2/HIDDA.forecasting_1.1.2.tar.gz", dependencies = TRUE)Alternatively, to install HIDDA.forecasting from the current sources, building the vignettes on your machine, you could use:
# install.packages("devtools")
devtools::install_github("HIDDA/forecasting", dependencies = TRUE, build_vignettes = TRUE)To browse the installed vignettes:
browseVignettes("HIDDA.forecasting")Note that the “CHILI” vignettes include source("setup.R"), which refers to a small R script available from the installed package at
system.file("doc", "setup.R", package = "HIDDA.forecasting")