ReCOVER
- Country and US state-level forecasts for COVID-19 using heterogeneous infection rate model - Data-driven identification of unreported case
This is a part of the following NSF project:
ReCOVER: Accurate Predictions and Resource Allocation for COVID-19 Epidemic Response
PIs: Viktor K. Prasanna (prasanna@usc.edu), Ajitesh Srivastava (ajiteshs@usc.edu)
University of Southern California
This repository contains some codes for our ongoing work on NSF-funded project on COVID-19 forecasting.
We use our own epidemic model called SI-kJalpha - Heterogeneous Infection Rate with Human Mobility.
For live script for forecasting, run: plot_gen.mlx
For detecting unreported cases use: daily_explore_unrep.mlx
Our relevant presentation: https://www.youtube.com/watch?v=ll6k8wlxOFo
Our paper on forecasting: https://arxiv.org/abs/2004.11372
Paper on detecting unreported cases: https://arxiv.org/abs/2006.02127
引用格式
Ajitesh Srivastava (2026). ReCOVER (https://ww2.mathworks.cn/matlabcentral/fileexchange/75281-recover), MATLAB Central File Exchange. 检索时间: .
| 版本 | 已发布 | 发行说明 | Action |
|---|---|---|---|
| 2.01 | Fixed a forecast lag |
||
| 2.0 | Added smoothing in forecasting. Also added possibility of detecting unreported cases |
||
| 1.1 | Improved hyperparameter search. Added pre-calculated hyper-parameters for various days in the past.
|
||
| 1.0.1 | Added some comments |
||
| 1.0.0 |
