Estimating the level of noise in time series using entropy methods – Andrea Scarciglia (Università di Pisa)


Aula O1 - Polo Fibonacci


Identification of underlying noise in time series is essential for applications ranging from physiology to econometrics, with relevance in all fields of data science. In this talk we present methods based on the computation of different notions of entropy, that can quantify the level of noise perturbing time series obtained from dynamical systems.

Torna in cima