RtEstim
| REF | Liu et al. (2024) |
| Docs | https://dajmcdon.github.io/rtestim/ |
| Github | |
| Last commit | |
| Installation | via devtools |
Brief description
Rtestim is a method that uses the renewal equation and a provided serial interval distribution to estimate \(R_t\). Distinct from other methods, it uses a frequentist approach with an L1 smoothing penalty which decreases computation time and allows for locally adaptive estimates. The method estimates confidence bands for \(R_t\) and incidence.
Methods
This package contains the following methods:
- Locally adaptive estimator using Poisson trend filtering
- L1 smoothing
- Cross validation to select tuning parameters for the smoother
This approach is similar to filtering
Assessment
| Features | |
| Ability to nowcast/forecast | No |
| Incorporates delay distributions | No |
| Estimates expected cases | Predicts based on estimated \(R_t\) |
| Communicates uncertainty | Yes |
| Validation | |
| Documentation of package methods | Yes |
| Documentation of package implementation | Yes |
Sample code
The package website has this helpful vignette.