Estimates the mean and standard deviation of the serial interval distribution from outbreak data using the Expectation-Maximization (EM) algorithm developed by Vink et al. (2014). The serial interval is defined as the time between symptom onset in a primary case and symptom onset in a secondary case infected by that primary case.
Usage
si_estim(
dat,
n = 50,
dist = "normal",
init = NULL,
tol = 1e-06,
n_starts = 1,
n_routes = 4L,
wind = 1
)Arguments
- dat
numeric vector; index case-to-case (ICC) intervals in days.
- n
integer; number of EM algorithm iterations to perform. Defaults to 50.
- dist
character; the assumed parametric family for the serial interval distribution. Must be either "normal" (default) or "gamma".
- init
numeric vector of length 2; initial values for the mean and standard deviation. If NULL (default), uses the sample mean and sample standard deviation.
- tol
numeric; convergence tolerance for the EM algorithm. Defaults to 1e-6.
- n_starts
integer; number of random restarts for the EM algorithm. Defaults to 1.
- n_routes
integer; number of transmission routes to model. Must be >= 2. Defaults to 4 (Co-Primary, Primary-Secondary, Primary-Tertiary, Primary-Quaternary). Increasing this allows modelling longer transmission chains.
- wind
The window censure interval
Value
A named list containing:
mean: Estimated mean of the serial interval distribution (days)sd: Estimated standard deviation of the serial interval distribution (days)wts: Numeric vector of estimated component weights. Length is 2*n_routes - 1 for normal distribution, n_routes for gamma distribution.converged: Logical indicating whether the algorithm converged.iterations: Integer indicating the number of iterations performed.loglik: Log-likelihood of the fitted model.n_restarts: Number of restarts performed.n_routes: Number of transmission routes used.
References
Vink MA, Bootsma MCJ, Wallinga J (2014). Serial intervals of respiratory infectious diseases: A systematic review and analysis. American Journal of Epidemiology, 180(9), 865-875. doi:10.1093/aje/kwu209
Examples
# Example 1: Basic usage with simulated data, default 4 routes
set.seed(123)
simulated_icc <- c(
rep(1, 20), rep(2, 25), rep(3, 15), rep(4, 8)
)
result <- si_estim(simulated_icc)
# \donttest{
# Example 2: Using 5 routes
result_5routes <- si_estim(simulated_icc, n_routes = 5)
# Example 3: Using gamma distribution with 3 routes
result_gamma <- si_estim(simulated_icc, dist = "gamma", n_routes = 3)
# }