Answers for "Does the Time Series Modeler - Expert Modeler need data transformation?"
https://developer.ibm.com/answers/questions/223887/does-the-time-series-modeler-expert-modeler-need-d.html
The latest answers for the question "Does the Time Series Modeler - Expert Modeler need data transformation?"Answer by LoicB
https://developer.ibm.com/answers/answers/223889/view.html
<p>
Thank you Kenneth!
</p>Fri, 18 Sep 2015 19:42:21 GMTLoicBAnswer by Kenneth Jensen (IBM)
https://developer.ibm.com/answers/answers/223888/view.html
<p>
The Time Series Expert Modeler in IBM SPSS Modeler is (as far as I know) the same algorithm that is available in IBM SPSS Statistics, so the two should have the same functionality.
</p>
<p>
In IBM SPSS Modeler, the "Expert Modeler" does consider a limited number of potential transformations of both the target series and any transfer function:
</p>
<ul>
<li>
Differencing
</li>
<li>
Square Root transformation
</li>
<li>
Natural Log transformation
</li>
</ul>
<blockquote>
<p>
If predictor variables are specified, the Expert Modeler selects for inclusion in ARIMA models those variables that have a statistically significant relationship with the dependent series. Model variables are transformed where appropriate using differencing and/or a square root or natural log transformation
</p>
<p>
<a href="http://www-01.ibm.com/support/knowledgecenter/SS3RA7_17.0.0/clementine/timeseriesnode_general.dita">http://www-01.ibm.com/support/knowledgecenter/SS3RA7_17.0.0/clementine/timeseriesnode_general.dita</a>
</p>
</blockquote>
<p>
While these are the most commonly used transformations in Time Series modeling, there might be situations or data series that would require other and less common transformations to be stationary or to increase their predictive power.
</p>
<p>
Kind Regards,
</p>
<p>
<strong>Kenneth A. Jensen</strong>, M.Sc.(Econ.) <br>
Senior Managing Consultant - Analytics Specialist<br>
IBM Performance Marketing - Analytics & Data Practice
</p>Fri, 18 Sep 2015 19:18:39 GMTKenneth Jensen (IBM)