Taking a Broad Sample of Time Series

The model learns how to understand those times series when it sees a lot of them. And i think one other interesting in about that is that there was a pure deploning model applied to time series, and it was successful. So it was like a major thing to kind of inspire people move more in that direction. We actually show that we can do a decomposition into trend and seasonality,. where trend is like quasi monotonic and seasonalityis quasi periodic. The model pretty ch is forced to output the parimeters of that trend and sesonality, and it still works.