|02 COVID mortality forecasting Project1 #701115|
An existing forecasting model is made up of multiple waves, each of which follows the shape of a skew-normal density function, with model fitting done with Bayesian MCMC. A challenge of working with this model is the large number of latent variables which are not Gaussian. The model will be extended to make spatio-temporal forecasts, and improved algorithms for handling waves of infections must be built. Methods for handling censored, aggregated, and age-specific death counts are also required.