Time-series Influenza Activity Forecast Model
This model estimates the individuals effects of nonpharmaceutical interventions (mask-wearing intervention and mobility-related nonpharmaceutical interventions) as well as the effects of SARS-CoV2 interference, and forecast the influenza activity under varying scenarios, including the scenarios with different assumptions on nonpharmaceutical interventions. It was used in two recent papers:
Han, S., Zhang, T., Lyu, Y. et al. The Incoming Influenza Season — China, the United Kingdom, and the United States, 2021–2022. China CDC Weekly 3(39):10.39-1045. (2021). http://weekly.chinacdc.cn/en/article/doi/10.46234/ccdcw2021.253.
Han, S., Zhang, T., Lyu, Y. et al. Influenza’s plummeting during the COVID-19 pandemic: the roles of mask-wearing, mobility change, and SARS-CoV-2 interference. Engineering (2022). Forthcoming.
For an example run of estimation of individual effects of mask-wearing intervention, mobility change, and SARS-CoV2 interference in the four regions, including nother China, southern China, England and the United States, please see "example_estimation/Run/run_example.sh", which uses the data from "preprocessed_data". For an example run of forecast the influenza activity in the four regions in the 2021-2022 influenza season, please see "example_forecast/Run/run_example.sh", which uses the data from "preprocessed_data". For an example run of pre-processing the raw data from the four regions, please see "preprocessing.py", which generates the intermediate dataset "preprocessed_data".
This model first estimates the individual effects of NPIs. It consists of two self-correcting regularized multiple regression models, both of which are dynamically trained and regularized using the LASSO method. Estimated effects are then used to predict influenza activity in the future.
• Python version 3.6.13.
• R version 4.0.3.
• scikit-learn 0.24.2 (Python)
• tidyverse(R)
• glue(R)
• xlsx(R)
• cowplot(R)
• scales(R)
• The main folder includes six subfolders: scripts, raw data, pre-processed data, example_estimation, example_forecast, result.
• scripts/preprocessing.py: Python script for the pre-processing operation. The script performs with the raw data and generates the pre-processed data.
• scripts/estimation_forecast.py: Python script for the main operation. The script estimates effects of one-week increase in NPIs and forecasts seasonal influenza activities using data from the folder "preprocessed data".
• scripts/calculate_plot_statistics.py: Python script to calculate mean, upper and lower confidence intervals of percent positivity at each time point.
• scripts/calculate_report_statistics.py: Python script to calculate mean, upper and lower confidence intervals of reduced percenty positivity within influenza seasons.
• scripts/param_mask_reg.py: Python script to select the parameters for the second muliple regression model that is used to estimate effects of a 1-week increase of mask-wearing interventions.
• scripts/auxiliary_name_file.py: Python auxiliary script to rename data files.
• scripts/auxiliary_gen_excel_data.py: Python auxiliaryscript to generate data as csv format for plotting in R.
• raw data: This folder contains the raw data for four study regions mentioned in the above two papers, nothern China, southern China, England and the USA.
• proprecessed data: This folder contains the intermediate data after pre-processing.
• example_estimation: The folder contains Python scripts for an example of running estimation model. The demo uses data from the four regions we studied in the paper. Output data are stored in the folder "result".
• example_forecast: The folder contains Python scripts for an example of running forecast model. The demo uses data from the four regions we studied in the paper. Output dataare stored in the folder "result".
• result: The folder contains output data generated by running examples of estimation and forecast models.
• example_estimation/Source data: The folder contains all generated data for reproducing the figures in the paper : Han, S., Zhang, T., Lyu, Y. et al. Influenza’s plummeting during the COVID-19 pandemic: the roles of mask-wearing, mobility change, and SARS-CoV-2 interference. Engineering (2022). Forthcoming.
• example_estimation/Source plot: The folder contains the plotting code for reproducing the figures in the paper : Han, S., Zhang, T., Lyu, Y. et al. Influenza’s plummeting during the COVID-19 pandemic: the roles of mask-wearing, mobility change, and SARS-CoV-2 interference. Engineering (2022). Forthcoming.
• example_forecast/Source data: The folder contains all generated data for reproducing the figures in the paper : Han, S., Zhang, T., Lyu, Y. et al. The Incoming Influenza Season — China, the United Kingdom, and the United States, 2021–2022. China CDC Weekly 3(39):10.39-1045. (2021). http://weekly.chinacdc.cn/en/article/doi/10.46234/ccdcw2021.253.
• example_forecast/Source plot: The folder contains the plotting code for reproducing the figures in the paper : Han, S., Zhang, T., Lyu, Y. et al. The Incoming Influenza Season — China, the United Kingdom, and the United States, 2021–2022. China CDC Weekly 3(39):10.39-1045. (2021). http://weekly.chinacdc.cn/en/article/doi/10.46234/ccdcw2021.253.