Michael DeWitt medewitt

Text once models can be pre-compiled

The first time you fit a model with {epinowcast} will be slow compared to future fits! This is normal as all of the Stan code is being compiled for the first time. Note that the compiled Stan binaries will live in your user package directory (or wherever you have defined your package libraries). Because of this package architecture

Each time you install a new version of {epinowcast} you will experience this longer compilation time the first time you fit a model.

The Shenzhen CDC identified 391 SARS-CoV-2 cases from January 14 to February 12, 2020 and 1286 close contacts. We compare cases identified through symptomatic surveillance and contact tracing, and estimate the time from symptom onset to confirmation, isolation, and hospitalization. We estimate metrics of disease transmission and analyze factors influencing transmission risk.

Cases were older than the general population (mean age 45) and balanced between males (187) and females (204). Ninety-one percent had mild or moderate clinical severity at initial assessment. Three have died, 225 have recovered (median time to recovery is 21 days). Cases were isolated on average 4.6 days after developing symptoms; contact tracing reduced this by

age_group	pct_sympto_hospital	pct_hosp_icu	infection_fatality_ratio
0-9	0.00	0.05	0.00
10-19	0.00	0.05	0.00
20-29	0.01	0.05	0.00
30-39	0.03	0.05	0.00
40-49	0.05	0.06	0.00
50-59	0.10	0.12	0.01
60-69	0.17	0.27	0.02
70-79	0.24	0.43	0.05
80+	0.27	0.71	0.09

	# Purpose: This is just a play model to think about time spent on large and big grants/opportunities and
	# overall likelihood of a successful result (undefined, just general feel).
	# 2080 is the hours in a year if working 40 hours (because everyone does that...)

	using JuMP
	using Ipopt
	model = Model(Ipopt.Optimizer)

	p1 = .05 # Probability of winning grant 1
	sx1 = .4 # Probability of successful work (class 1)

	library(tidyverse)
	n.sim <- 1000
	lambda <- 8/45 # 8 nursing homes in 45 days


	set.seed(336)
	n.sim <- 1000
	mu <- 20 # 20 hospitalizations mean
	sigma <- 5 ## =/- 10 patients
	sigma.sq <- sigma^2

	library(data.table)
	library(tidyverse)

	path_seirds_model <- system.file("examples/discrete_stochastic_seirds.R", package = "odin")

	seirds_generator <- odin::odin(path_seirds_model)

	x <- seirds_generator()

	x_res <- as.data.frame(replicate(100, x$run(0:365)[, -1]))

	#' Stan Model Base General
	#'
	#' @param dat the dat sent to the model
	#' @param iter 2400
	#' @param warmup 1000
	#' @param chains 4
	#' @param control default controls
	#' @param ... Arguments passed to `rstan::sampling` (e.g. iter, chains).
	#' @return An object of class `stanfit` returned by `rstan::sampling`
	#' @export