Predictive Analytics - as Commodity Off The Shelf?

[Former Member]

Hi,

According to Gartners Hype Cycle for Emerging Technologies Predictive Analytics is now firmly in the Plateau of Productivity. Meaning that mainstream adoption starts to take off. Criteria for assessing provider viability are more clearly defined. The technology’s broad market applicability and relevance are clearly paying off.

I would say that this is very much in alignment with the tool capabilities that SAP Predictive Analysis brings to the market. Not only is it in my view much simpler to use than "R" or some of the other big statistical packages out there but it is also delivers a high user convenience. With the possiblity of "drag & drop" statistical algorithms on to your data source and uncover previously hidden information and patterns.

Moreover if you already have a productive BusinessObjects universe-based solution SAP Predictive Analysis can very quickly be applied and simply re-use that investment. This should allow for quick and seamless implementation. The very same argument goes for companies with HANA in-memory technology - but in with the ability for much faster and more data. And not forgetting packaging the build algorithm into user applications for real-time usage scenarios.

Gartner hype cycle 2013. Cycle for Emerging Technologies Maps Out Evolving Relationship Between Humans and machines, published 19th august 2013. Reference: http://www.gartner.com/newsroom/id/2575515

Best regards,

Kurt Holst

References:

Gartner hype cycle 2013. Cycle for Emerging Technologies Maps Out Evolving Relationship Between Humans and Machines: http://www.gartner.com/newsroom/id/2575515

Gartner actually uses two different definitions for predictive and prescriptive analytics:

Predictive analytics answers the question what will happen. This is when historical performance data is combined with rules, algorithms, and occasionally external data to determine the probable future outcome of an event or a likelihood of a situation occurring.

Prescriptive analytics: predicting future outcomes by also suggesting actions to benefit from the predictions and showing the implications of each decision option.

Wikipedia:

http://en.wikipedia.org/wiki/Prescriptive_analytics

http://en.wikipedia.org/wiki/Hype_cycle

[RafkeMagic]

I agree in that it's simpler to use...

However, the error messages are extremely unclear, so are some of the prerequisites for certain algorithms.

It's good that it allows one to create ones own "R" algorithms though ...

[Former Member]

Hi Raf,

Thanks for your comment.

You are absolutely right the capability of using custom built R functions in SAP Predictive Analysis is enhancing the functionality to a completely new and better level.

Regarding your second point on error messages. I am sure you know this but SAP PA generates a logfil that is default stored in %TEMP%\sapvi\logs

This logfile shows in more detail what is going on and could also assist support if you create an OSS message.

If you want need more information you can change the logging details as described below (from Vishwa' blog on tracing):

1. Create a folder in any location for generating logs.

Note: Ensure that you have "write" permission to the folder. For example, C:\logs.

2. Create the BO_Trace.ini file and add the following trace details to it.

active=false;

severity='E';

importance=xs;

size=1000000;

keep_num=437;

alert=true;

Save and close the BO_trace.ini file.

4. Place the BO_Trace.ini file under C:\logs.

5. Set up the following environment variables:

• BO_TRACE_LOGDIR = C:/logs

• BO_TRACE_CONFIGDIR = C:/logs

• BO_TRACE_CONFIGFILE = C:/logs/BO_Trace.ini

6. Restart the application.

Hope this helps you out.

Best regards,

Kurt Holst

[RafkeMagic]

Thx, Kurt!

I'll try your suggestion on Monday (on a BW project at a client today)... the "default" logfiles do indeed contain more information, however most of it is meaningless (when it comes down to the actual error). At least in my experience...