I have wonder about exponential smoothing

For econometric discussions not necessarily related to EViews.

Moderators: EViews Gareth, EViews Moderator

hussam
Posts: 6
Joined: Mon May 15, 2017 3:40 am

I have wonder about exponential smoothing

Postby hussam » Mon Jul 10, 2017 4:56 pm

Hi, I have a request for clarification about this procedure ::
i used exponential smoothing with model specification(MNN) in ETS Smoothing of dowjones returns series(200 obs), then ploted two series (original series and smoothed series) as group and noted the following : there is one lag between these series, so i deleted a first observation from smoothed series and shift above the next observations and identify with original series via garph and find a fairly acceptable result.
then i made two residual series (returns-smoothed series) , (returns-smoothed series with deleted first obs) for estimate garch(1,1) model and found the following: both of estimated coefficients of two models are identity but only AIC ,BIC and Logl are different , where the AIC ,BIC for garch model of series with deleted first obs are better .
my question :: Is my previous method correct? and why?

i attached a workfile for apply
note return : original series,
return_sm_new : smoothed series with deleted first obs
Attachments
dow.WF1
(124.32 KiB) Downloaded 223 times

Return to “Econometric Discussions”

Who is online

Users browsing this forum: No registered users and 19 guests