The easiest way to improve time series forecasting?

by Mr Carl Remlinger

Wednesday 15 Mar 2023, 11:30 → 13:00 Europe/Zurich

OHSA/E13 (https://psich.zoom.us/j/64271153417?pwd=eFFqRDY4UmZhYUh3SFlRa1dGanJ5UT09 Meeting-ID: 642 7115 3417 Kenncode: 502508 )

Abstract: 
Machine learning algorithms dedicated to time series forecasting have gained a lot of interest over the last few years. One difficulty lies in the choice between several algorithms, as their estimation accuracy may be unstable over time.  Online Aggregation of Experts combines a finite set of forecasting models without making assumptions about the data. The mixture is updated continuously when data becomes available. This is a desirable feature in non-stationary environments as it allows to reconsider at each time step what are the best estimators. 
 

REGISTRATION DEADLINE ==> MO 13/03/23, 12 am