I have several questions regarding the ETS model in statsmodels library. The description of the model can be found here.
- The default
initialization_methodis estimated. In the description, it says ‘estimated’ uses the same heuristic as initial guesses, but then estimates the initial states as part of the fitting process. What does this mean? What is the heuristic value as the initial guesses? How does the estimation work? Does it try to minimize the sum of squared errors of the one-step ahead forecast? - If I specify
damped_trend = True, how does the model choose the optimal damping parameter?
