Exponential Smoothing using ets function

I am trying to build an exponential smoothing model and wanted to know what does the "ETS(M,A,N)" means, in terms of what does M, A and N stand for and what does it tell about this model. I also wanted to know if through this output there is any way to tell the model output is Additive or Multiplicative. Lastly, I wanted to know if through this output I can know the model is factored trend and seasonality? Please let me know if you can help me understand the output and if not, I can use any other functions to understand these items! Thank you for your help.


Call: ets(y = ts1)

Smoothing parameters: alpha = 1e-04 beta = 1e-04

Initial states: l = 344613.9475 b = 9014.0389

sigma: 0.5051

 AIC     AICc      BIC 

1272.851 1274.430 1281.772

M is for multiplicative, A is for additive and N is for none. In the ETS notation, the first parameter is for the Error, the second for Trend and the third for Seasonality. The ets() function selected a model with multiplicative error, additive trend and no seasonality. The model thus produced smoothing parameters for the level (alpha) and trend (beta) but not for seasonality (gamma).

I suggest you read Chapter 8 on Exponential Smoothing in the excellent book by Hyndman and Anthanasopoulos at otexts.com/fpp3/

1 Like

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.