DAK-Projects/COVID19-SEIRD-model-for-Africa-2021

# COVID19-SEIRD-model-for-Africa-2021 This model analyses data for dynamic modelling of the true burden of COVID-19 and deaths in the African region. This being country specific modeling, the burden in each of the countries is computed using the Partially Observed Markov Processes (POMP). The software used is pomp King et al., 2016 and the data is also attached herein. #Get started 1. Create a working directory, which may be local or online 2. Remove any object that could be in the memory of the R server 3. Create a data vector containing each of the 47 countries of interest in Africa. This will be used for looping (from country to country) 4. Load the packages needed, and set the random seed, to allow reproducibility 5. Select the countries and load the two datasets: CovCasesWeekly.txt that contains the COVID-19 reported cases and deaths; DynCovInitGA.txt that contains the initialization values for all the parameters used in the model. 6. Filter the data by each of the countries and fit the model process #7. In the modelling phase: a. Fit cases and deaths data to obtained the measurement model parameters for the distribution assumed b. Initializing the parameters c. Derive immunization rates proportions of the population at any time t, using a logistic growth curve d. Obtain the seasonality in the data by fitting a Fourier transform to the observed data. The parameters form this model are used to introduce the waves in the force of infection rate at the same points as in the observed data e. Fit the measurement model C snippets by defining random simulator of measurement model with the parameters obtained from the data f. Initialize and define the process as per the state-space model developed g. Simulate and optimize the model based on the initialization parameters provided h. Aggregate the total number of persons under either sub-model, that is, immunized and not immunized arms and get the 95% confidence intervals i. Codes for plotting the same are als …