spell check

R/create.R

@@ -345,7 +345,7 @@ create_gp_data <- function(gp = gp_opts(), data) {
 #' # Applying a observation scaling to the data
 #' create_obs_model(obs_opts(scale = list(mean = 0.4, sd = 0.01)), dates = dates)
 #'
-#' # Apply a custom week week lenght
+#' # Apply a custom week week length
 #' create_obs_model(obs_opts(week_length = 3), dates = dates)
 create_obs_model <- function(obs = obs_opts(), dates) {
   data <- list(

inst/WORDLIST

@@ -1,285 +1,107 @@
-abbott
-ac
-accum
-adj
-al
-args
-aut
-backcalc
-backcalculation
-badr
-Badr
-behavior
-BH
-Biarch
-bool
-Bosse
-bp
-BugReports
-ccdf
-cdf
-centered
-centred
+BMC
+Bracher
 CMD
-cmf
 CMF
+CMMID
+COVID
+Centers
+CoV
 Codecov
-com
-const
-conv
-convolving
-countrycode
-coutn
+Cori
 Covid
-covr
-cowplot
-cre
-crtp
+Dirichlet
+Dockerfile
+ECDC
+EpiNow
+Epinow
+Gruson
+Habakuk
+Hain
+Hamada
+IO
+Inf
+Jit
+Kanjilal
+Kucharski
+LN
+Leeuwen
+Lifecycle
+Lipsitch
+Lison
+LogNormal
+Makevars
+Matern
+Mayol
+Monkeypox
+NUTs
+Nowcasts
+ORCID
+PLoS
+PMF
+PRs
+Poisson
+Quilty
+Riutort
+Rmarkdown
+Rstan
+Rtools
+Runtimes
+SARS
+Sanjat
+Wellcome
+backcalc
+backcalculation
+com
 csv
-ctb
-ctor
 dd
-de
-deconvolution
 defato
-dev
-dewitt
-DeWitt
-diagSPD
-dimss
 discretised
 dist
 doi
-dplyr
-dt_settings
-data.table
-ECDC
-ecuyer
-edu
-Eigen
-ellis
-elt
-endif
+dt
 epidemia
 epiforecasts
-EpiForecasts
 epinow
-Epinow
-EpiNow
-EpiSoon
-eq
-et
-fi
+etc
 filepath
-finalising
-focusses
-focussing
-forecastHybrid
 frac
-ggplot
 github
-gmail
-Gostic
-gov
 gp
-gqs
-gt
-Hamada
-hellewell
-Hellewell
-hickson
-Hickson
-hpp
-HPP
-ht
-https
-ifndef
-inf
-Inf
-initialisation
-initialise
-inits
-int
-io
-IO
-istream
-jacobian
-james
-jhu
-joel
-joseph
-jr
-kableExtra
-katharine
-kgostic
-knitr
-lang
-LazyData
-lgamma
-lifecycle
-Lifecycle
-lik
+lengthscale
 lineranges
-LinkingTo
 logmean
 lognorm
 logsd
-lp
-lpmf
-lshtm
-lt
-lte
-lub
-lubridate
-magrittr
-Makevars
 matern
-Matern
-math
-Mayol
+mathcal
+mathrm
 mc
-Meakin
-meaking
-meanlog
-mee
-Mee
 metacran
-metoffice
-mf
-misspecified
-monticone
-Monticone
-msg
-munday
-Munday
 negbin
-nht
-Nikos
-normalisation
 nowcast
+nowcasting
 nowcasts
-Nowcasts
-num
-NUTs
-nz
-oberserved
 obs
-optimisation
-Optimised
-oR
-ORCID
-org
-ostream
-ot
 parallisation
-param
-params
-paul
-periodicty
-pietro
-Pietro
-pmf
-PMF
-pmfs
-poisson
-Poisson
-pos
-pow
-precommit
-priori
-prob
-prog
+parameterised
+pngs
 progressr
-propto
-PRs
-pstream
-purrr
-Rcpp
-RcppEigen
 rds
-README
 realland
-Rebased
-Recoded
-recon
-reprex
-rethrow
-Riutort
-rlang
-rmarkdown
-Rmarkdown
-rnaturalearth
-rng
-roxygen
-Roxygen
-RoxygenNote
 rstan
-Rstan
-rstaninc
-rstantools
-Rtools
-Rts
-Runtimes
-rvalue
+runtimes
 sam
 sd
-sdlog
 se
-sebastian
 seealso
-sherratt
-Sherratt
 sim
-sophie
-spd
-SPD
-sR
 stan
-standardised
 stanfit
-StanHeaders
-std
-str
-stringstream
-struct
-styler
-summarise
-summarises
-supress
-SystemRequirements
 testland
-testthat
-thompson
-thresold
 tibbles
-tidyr
-tidyverse
 timelimit
-timeseries
-tparams
-trunc
-truncnorm
-typedef
-typename
-uchicago
-uk
-unconstrain
-unito
-unobs
-uot
-util
-val
-vals
-var
-VAR
-variational
-Variational
-vars
 vb
-vec
-vectorised
+vscode
 warmup
 wellcomeopenres
-wl
-www
 yyyy

vignettes/estimate_infections.Rmd

@@ -61,7 +61,7 @@ This model used the following priors for the observation model
 
 $\phi$ is truncated to be greater than 0 and with $\xi$, and $w$ normalised to sum to 1. Other priors are set by the user.
 
-$GP_t$ is an approximate Hilbert space Gaussian process as defined in [@approxGP] using a Matern 3/2 kernel with a default boundary factor of 1.5 and basis functions scaled to be 20% of the number of days fitted. The length scale of the Gaussian process was given a log-normal prior with a mean of 21 days, and a standard deviation of 7 days truncated to be greater than 3 days and less than the lenght of the time-series. The standard deviation of magnitude of the Gaussian process was assumed to be 0.1. These settings are all changeable by the user. In addition the user can opt to make use of a different generative process or to instead remove the dependency on the previous value of $R_t$ each of these options impacts run-time and may alter the best use-case for the model.
+$GP_t$ is an approximate Hilbert space Gaussian process as defined in [@approxGP] using a Matern 3/2 kernel with a default boundary factor of 1.5 and basis functions scaled to be 20% of the number of days fitted. The length scale of the Gaussian process was given a log-normal prior with a mean of 21 days, and a standard deviation of 7 days truncated to be greater than 3 days and less than the length of the time-series. The standard deviation of magnitude of the Gaussian process was assumed to be 0.1. These settings are all changeable by the user. In addition the user can opt to make use of a different generative process or to instead remove the dependency on the previous value of $R_t$ each of these options impacts run-time and may alter the best use-case for the model.
 
 ### Beyond the forecast horizon
 

vignettes/estimate_secondary.Rmd

@@ -17,9 +17,9 @@ knitr::opts_chunk$set(
 **This is a work in progress. Please consider submitting a PR to improve it.** 
 
 This model is based on a discrete convolution of primary cases, scaled based on the fraction (here described as the secondary fraction but depending on application potentially being the case fatality ratio, case hospitalisation ratio, or the hospitalisation fatality ratio), and a delay distribution that is assumed to follow a discretised daily log normal distribution.
-This model can be thought of as a disrete time ordinary differential equation approximation generalised to log normal, rather than exponential, delay distributions.
+This model can be thought of as a discrete time ordinary differential equation approximation generalised to log normal, rather than exponential, delay distributions.
 
-We generalise this simple model beyond the incidence cases to also include prevalence indicators (for example hospital admissions and occupancy) where the secondary notifications can be thought of as depending on secondary notifications from the previous timestep, scaled current primary notications, and minus scaled historic primary notifications weighted by some delay distribution. 
+We generalise this simple model beyond the incidence cases to also include prevalence indicators (for example hospital admissions and occupancy) where the secondary notifications can be thought of as depending on secondary notifications from the previous timestep, scaled current primary notifications, and minus scaled historic primary notifications weighted by some delay distribution. 
 
 This model can be defined as follows,
 
@@ -29,7 +29,7 @@ This model can be defined as follows,
 
 where $S_t$ and $P_t$ are observed primary and secondary notifications, $\hat{S}_t$ are expected secondary notifications, $\delta_p = 1$ and $\delta_c = -1$ when $S_t$ is a prevalence measure, $delta_p = 0$ and $\delta_c = 1$ when it is an incidence based measure.
 $\alpha$ and $\xi$ are defined as the secondary fraction and delay from primary to secondary notification (or delay from secondary notification to recovery etc in the prevalence case) with $\alpha$ typically being of most interest to those interpreting the models posterior estimates.
-We further assume that $\xi$ follows a discretised log normal distibution described by its mean $\mu$ and standard deviation $\sigma$ on the log scale (where we take the cumulative mass function for time $t$ minus the cumlative mass function for $t-1$) normalised by the maximum allowed delay $D$ such that $\sum^D_{\tau=0}{ \xi(\tau | \mu, \sigma)}  = 1$.
+We further assume that $\xi$ follows a discretised log normal distribution described by its mean $\mu$ and standard deviation $\sigma$ on the log scale (where we take the cumulative mass function for time $t$ minus the cumulative mass function for $t-1$) normalised by the maximum allowed delay $D$ such that $\sum^D_{\tau=0}{ \xi(\tau | \mu, \sigma)}  = 1$.
 
 The above definition captures our mechanistic assumptions for the expectation of secondary notifications but does not account for potential observation noise or reporting patterns.
 Here we assume a negative binomial observation model (though our implementation also supports a Poisson observation model) in order to capture potential reporting overdispersion $\phi$ and adjust expected counts using an optional day of the week effect based on a simplex $\omega_{(t \mod 7)}$ (such that $\sum^6_{t=0}{w_t} = 7$ so the total effect over a week is balanced).