We implements approaches such as:
Dynamic Time Series Model applying State Space modelling, Bayesian statistics and Kalman filter.
Logit Multinomial PLS (Partial Least Square) to improve robustness of multinomial models.
MCMC (Monte Carlo Markov Chain), which allows to estimate "posterior" expectations in complex problem.
Agent Based Model which allows to reproduce characteristics of real outbreaks and make their predictions easy to understand.