Time series models with time varying coefficients: applications in Energy, Meteorology, Cyber ​​Security, Insurance and Retail.

Cristiano Fernandes and Collaborators

We use the recently proposed class of time series models with time varying coefficients named dynamic score models to address relevant practical problems in Energy (prediction of wind power capacity factors), Meteorology (prediction of extreme values ​​of rainfall series), Insurance (reserve estimation IBNR, prediction of mortality rates, collective risk estimation), Cyber ​​Security (detection of anomalous patterns in access to websites) and Retail (sales prediction under promotions and price changes).