EstmatingRtDeepL

Codes for the paper " Estimating time-varying reproduction number by deep learning techniques " submitted to JAAC.

Note:

• In 2023/05/20, I updated project.toml file to use Julia 1.9

• In 2023/06/26, I updated project.toml and logistic.jl, one can use

• In 2023/06/26, I updated project.toml and logistic.jl as a template and modify other files such as mediaimpact.jl.

Document Descriptions

• Toy Models: logisticgrowth.jl, subexpotential.jl, mediaimpact.jl
• DeepLearningEffectiveReproductionNumber Estimating effective reproduction number of Ontario first wave data.
• Rt_Methods_Comparison Kalman, EpiNow2 and EpiEstim Methods.
• Data Summarization: comparison_emsemble.jl

Instructions for running the code

Part One: Estimating time-varing reproduction number by universal differential equations

Julia Language:

• Step 1: Download Julia and Configuration Julia Environment. Download Julia One can search online for how to configure the julia environment.

• Step 2: Git Clone this repo or download the document.

• Step 3: cd to the repo folder.

using Pkg
Pkg.instantiate(".")

Then many packages will be downloaded. My project includes many packages one may not use. You can also set up your personal project environments following the guide: Introduction · Pkg.jl DifferentialEquations.jl,DiffEqFlux.jl, Plots ,DataFrames.jl, CSV.jl and Flux.jl are necessary.

• Step 4: Run the codes.

If one are not familiar with Julia, one can see for more details in Julia Documents: Julia Documentation · The Julia Language

Part Two: Comparisons and ensemble methods: EpiEstim, EpiNow2 and Deep Learning

In folder "Rt_Methods_Comparison".

R Language.

One need to configure the path environments in the codes.