Page **1** of **1**

### ExpSmooth (Exponential Smoothing - requires R)

Posted: **Mon Apr 05, 2010 1:36 pm**

by **EViews Gareth**

This thread is about the

ExpSmooth Add-in that Performs an expanded set of exponential smoothing and forecasting techniques, including automatic model selection. Note R is required for this add-in.

**Change Log**- 2010/03/24 - Initial release
- 2010/04/09 - Fix for bug caused by the global option setting for sending variables over to R

### Re: ExpSmooth (Exponential Smoothing - requires R)

Posted: **Fri Apr 09, 2010 2:31 pm**

by **EViews Gareth**

New update posted, fixing an issue where users who had changed their global options to send over series to R in Upper Case would have problems.

### Re: ExpSmooth (Exponential Smoothing - requires R)

Posted: **Tue Sep 28, 2010 4:01 am**

by **Stefano**

Hi all!!

I am trying to use ExpSmooth Add-in, I installed both R and the Forecast package but now when I am trying to run the add-in from the time-series object in Eviews 7 I get an error message like "R is not installed". Could you please help me to solve this problem?

Thank you in advance

### Re: ExpSmooth (Exponential Smoothing - requires R)

Posted: **Tue Sep 28, 2010 8:06 am**

by **EViews Gareth**

Do you have the statconnDCOM package installed? (See the EViews Command Reference, Chapter 7, for more details).

### Re: ExpSmooth (Exponential Smoothing - requires R)

Posted: **Tue Sep 28, 2010 8:36 am**

by **Stefano**

Thank you Gareth,

I didn't know I need to install it...I will immediately do it

Regards

### Re: ExpSmooth (Exponential Smoothing - requires R)

Posted: **Thu Dec 02, 2010 12:54 am**

by **basher**

Hello,

I installed R, installed statconnDCOM, intalled also Scilab 5.2. After installing "rscproxy", when I verify the installation by clicking "Server 01 - Basic test", I receive the following message:

Loading StatConnector Server... Done

Initializing R...Function call failed

Code: -2147221485

Text: installation problem: unable to load connector

Releasing StatConnector Server...Done

May be I should ask this question to "R", thought I might get better response from this platform. Thank you.

### Re: ExpSmooth (Exponential Smoothing - requires R)

Posted: **Thu Dec 02, 2010 8:27 am**

by **EViews Steve**

Hi basher:

My guess is that some of your components are not compatible with each other. Instead of installing each item separately, I would use the installer that installs everything for you in one go.

If you go to this page:

http://rcom.univie.ac.at/download.htmlYou can download a file called "RAndFriendsSetup2120V3.1-11-1.exe". I believe this will install everything with known compatible versions to get the statconnDCOM working. (You probably have to uninstall everything else first).

Steve

### Re: ExpSmooth (Exponential Smoothing - requires R)

Posted: **Thu Dec 02, 2010 10:11 am**

by **basher**

Thank you Steve. It is working well.

### Re: ExpSmooth (Exponential Smoothing - requires R)

Posted: **Thu Dec 02, 2010 10:23 am**

by **basher**

While the "expsmooth@ add-in allows to get estimates from 15 different models, I was wondering whether it is possible to integrate the state space innovations (within Eviews) in a future update of this very helpful routine. This idea is explained in the accompanying paper of the "forecast" package. Thank you.

### Re: ExpSmooth (Exponential Smoothing - requires R)

Posted: **Mon Mar 12, 2012 12:06 am**

by **jaekwan7337**

Hi,

I am currently using Exponential Smooth Transition model to test a existence of bubble. It is nonlinear model, and I read one of your comments here that this add in will be allow me to get more accurate estimate for the coefficient, as I dont know how to do grid search from eviews.

the model is as below.

DPQUK=C(1)*PQUK(-1)*(1-EXP(-C(2)*PQUK(-1)^2))+DPQUK(-1)

I have installed the add in, but the add in only works with the single series, rather than the model estimation.

Can you give me some advice?

cheers,

Jake

### Re: ExpSmooth (Exponential Smoothing - requires R)

Posted: **Mon Mar 12, 2012 12:47 am**

by **trubador**

jaekwan7337 wrote:I have installed the add in, but the add in only works with the single series, rather than the model estimation.

That is because it falls into the category of "smooth transition" models, which is a very different concept than smoothing. Following discussion will be helpful:

viewtopic.php?f=4&t=3712