Estimate the Effective Reproductive Number with Trend Filtering
Add confidence bands to estimated Rt or incidence curves
Rt estimation algorithm configuration
Leave-kth-out cross validation for choosing a optimal parameter lambda
Calculate the total infectiousness at each observed time point.
Compute the discretized density function for gamma distribution
Estimate Rt using smoothness-penalized Poisson likelihood
Fitted cv_poisson_rt
Interpolate (or extrapolate) Rt estimates to intermediate design point...
Plot cv_poisson_rt
Plot estimated Rt values from a poisson_rt object
Plot estimated confidence bands for an estimate of Rt
Predict observed data using estimated Rt
Predict observed data using estimated Rt
Use trend filtering, a type of regularized nonparametric regression, to estimate the instantaneous reproduction number, also called Rt. This value roughly says how many new infections will result from each new infection today. Values larger than 1 indicate that an epidemic is growing while those less than 1 indicate decline. For more details about this methodology, see Liu, Cai, Gustafson, and McDonald (2024) <doi:10.1371/journal.pcbi.1012324>.
Useful links