the 3 input arrays have different length, what's the semantics of putting them together to pass as in as input? Are they:
-- represents the same time window? From the np.sin call with the same start and end (0, 20) (and this model only do univariate time series forecasting), guess they are. But then why the high frequency 0 has fewer data points (100) than the low frequency 2 (400)? This is confusing. Should the frequency_input array be reversed?
-- or different time windows, but end aligned?
-- or different time windows, but appended together one by one (in time)?
In this code:
the 3 input arrays have different length, what's the semantics of putting them together to pass as in as input? Are they:
-- represents the same time window? From the
np.sin
call with the same start and end(0, 20)
(and this model only do univariate time series forecasting), guess they are. But then why the high frequency 0 has fewer data points (100) than the low frequency 2 (400)? This is confusing. Should thefrequency_input
array be reversed?-- or different time windows, but end aligned?
-- or different time windows, but appended together one by one (in time)?
Thanks.