My agenda is finding patterns and possible a model that describe my data.
My data is comprised of multiple time-series uni-variate sets.
data_set = [pd.Series(np.random.normal(size=10), range(10)),
pd.Series(np.random.normal(size=10), range(10)),
pd.Series(np.random.normal(size=10), range(10)),...]
Each of these series follows some time-series assumptions - meaning there is a dependency between X and X-1, X-2 .. at each series.
All of the series represent similar process, so they have some-what similar properties(value scale, change points, seasonality) but they are not identical.
As to my understanding all time-series models are good for learning one time-series with no implication on other series. Is there a way for fitting a model that is best for all series? What is the method for learning/model across multiple related time series?
