Discussion: View Thread

I have project coming up invoving forecasting and where the computing environment is MS Sequel Server.  They want turn key type tools developed using MS Sequel Server Time Series module.  This is a business environment where they are not interested in SAS or other more statistical environs.  Does anyone have experience with this that they can comment on.

The basic job is to forecast a certain variable of interest 3 to 6 months in advance.  There is no lack of data just this limitation on tools used. Part of what I am referencing is described at http://msdn.microsoft.com/en-us/library/bb677216.aspx

I have not used this email system much but did send out this request a few weeks ago but was not sure where it went so am redoing it.

Thanks in advance for any help on this.


-------------------------------------------
J. Dobbins
-------------------------------------------

Hello Mr. Dobbins,

I've used MS SQL Server 2005 long time ago, Oracle a few years ago, and just finished my graduate course in Time Series Analysis (mainly using R or Minitab). So, specifically, I have not used the ARIMA function in the MSSQL environment, but I have some thoughts and could try to help with this. 

In your first email you asked whether MS SQL was user-friendly. Is your concern more oriented to the development environment, or to the use of the MS SQL ARIMA/ARTXP algorithm itself? MS SQL has the SQL Server Management Studio (SSMS) development/management environment, which in some sense is similar to Oracle SQL Developer. They both have reporting tools, but I have not used them (just worked on raw queries/scripts).

Regarding the ARIMA algorithm, the documentation seems to be quite useful. The first thing to find out would be what version of SQL Server the business institution is running. It seems SQL Server 2005 used the ARTXP algorithm, SQL Server 2008 incorporated the ARIMA algorithm, and its customization features may vary between SQL Server 2008 and 2012. So, the flexibility on using the algorithm depends on the SQL Server version, and this will determine which parameters you can use while working with the ARIMA/ARTXP model. 

Best regards,

--------------------------------------------------------------
Mariana Varotto
Penn State University - Master of Applied Statistics Program
--------------------------------------------------------------