2
$\begingroup$

I am trying figure out what is the difference between Brown's linear model for double exponential smoothing and Holt's model. So the differences can be implemented into a Holt model using if statements. Aside from having alpha for both the trend and smoothing component.

I am having a problem where the forecasting occurs when the level equation is added to the trend equation to forecast:

IBM SPSS show that when forecasting k steps ahead ((k-1)+(alpha^-1)) is used.

But in other books and articles it shows that for the forecast equation for the level and trend is changed to 'at' and 'bt' before being added together.

Are the SPSS equation and the other one the same? (I am asking because it is easier to implement one over the other)

Also what is recommended for the initial points? And did I miss any differences?

Improve this question
$\endgroup$
5
  • $\begingroup$ Another difference than that in Holt's there are two parameters? Please, be more specific, your question is somewhat unclear $\endgroup$
    – Germaniawerks
    Commented Jan 3, 2014 at 22:17
  • $\begingroup$ @Germaniawerks mentioned that it is using alpha for both components $\endgroup$
    – user3084006
    Commented Jan 4, 2014 at 0:26
  • $\begingroup$ @Glen_b That is just to show the differences between both articles $\endgroup$
    – user3084006
    Commented Jan 4, 2014 at 5:54
  • $\begingroup$ @Glen_b How about this to simplify the question "What is the forecast Equation for the Brown Linear Smoothing?" $\endgroup$
    – user3084006
    Commented Jan 4, 2014 at 6:20
  • $\begingroup$ Your edits now make sense. I wonder if your edits weren't showing before or something, but +1 now for sticking with it and making a better question. Both approaches are defined in the relevant Wikipedia article. They are not the same. $\endgroup$
    – Glen_b
    Commented Jan 4, 2014 at 8:05

0

Reset to default

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Browse other questions tagged or ask your own question.