The "ExponentialSmoothing" function in Python's "statsmodels.tsa.holtwinters" library gives you the option to set
trend equal to
multiplicative but it also lets you set a
seasonal parameter equal to
multiplicative. I believe the
trend is additive when the peaks & troughs of the time series remain roughly the same. Is this correct? How do I determine if the seasonal parameter is additive or multiplicative?
Here is my time series plotted along with its decomposed plots: