Complete list of Nodes in SAS Enterprise Miner

[deleted_user]

Hi,

I want to get the complete list of Nodes with some details in SAS Enterprise Miner 4.3 / 5.2. Can any one help me?

Thanks!

[ChrisHemedinger]

Hi,

This is a bit off-topic for this forum, which is meant for Enterprise Guide.

You might find some of what you need at:

http://support.sas.com/software/91x/emguiwhatsnew900.htm
and
http://support.sas.com/software/91x/emgui52whatsnew900.htm

Chris

[deleted_user]

I know that Enterprise Miner is "off-topic" for this forum, but are there plans to start a forum for Enterprise Miner/Text Miner?

[deleted_user]

Agree

[deleted_user]

> I know that Enterprise Miner is "off-topic" for this
> forum, but are there plans to start a forum for
> Enterprise Miner/Text Miner?

Where do we ask Enterprise Miner related questions?

[deleted_user]

Why not have the enterprise miner nodes as an add-on in enterprise guide, so that the functionality can be used in EG?

I am doing most of the data manipulation work before the predictive modeling in enterprise guide. Because I think it is easier to it in EG, but when I have the final data set suitable for modeling I need to use EM for the more advanced modeling stuff.

[JasonXin]

My two cents about EG and EM

The comment towards EG integrated with EM in one place is one stop in a long imagination: what happens after the two get married? Soon some may say why don't we just move Forecasting Sever or Marketing Optimization. I am not saying we should NOT think that way. Just perhaps a bit understanding why the two are separate may be more helpful?

1. SAS, like many other analytics product vendors, serves wide varieties of users in the market place. Some are more specialist. Some need to access broader sets of tools across, a bit of many. So user preference is one thing
2. If you set aside EG GUI and EM GUI, and focus on PROCS behind the two, you really are doing the two in one place. Obviously the choice of the place is EG. For example, you run EM models in batch. You can finish design a prototype. And then run faster and more varieties in EG. Behind each EM flow there is a SAS code flow produced that you can move to EG to run. The fact is the latest EG produces stunning high definition powered by new graph procedure like SGPLOT. It has been made compatible with ETS, STATS....
3. The bridging to conduct predictive modeling from EG already exists inside EG. If you have Add-on-for Microsoft (surely this does not work on Linux yet), you can build models using Rapid Predictive Model (RPM). Then use EM to point to the location of the model (just a XML file) for modification. I have seen many new and experienced EM users learn SAS EM tricks in this fashion. For a new EM user, it is timing consuming to figure to which node goes where and follow what.... Within RPM, you can select basic, intermediate and advanced to suite yourself
4. The latest high performance initiatives (HP) are leaning towards more integration. If you run SAS Visual Analytics (VA), for example, you begin with a HUB where all your licensed HP applications are listed there. More and more new HP products are being put out there as we sort out different infrastructures supporting HP. Integration based on in-memory, near real time technologies perhaps makes more sense.
5. Increasingly there are  intrinsic, analytic needs to mix among analytical packages. For example, when EM models are built using more and more transaction, time series inevitably will enter the practice, since transactions, by definition, have time stamps. Your EDA becomes more and more time wise. So, it appears that if some closer integration is desired or warranted, perhaps should first happen among analytical pieces.

What else?

Best

Jason Xin

[jakarman]

Em 4.2 / 5.2 that is old history as being SAS 8 or the first 9  release.  With SAS 9.1.3 EM 5.3 and eguide 4.1 was the combined numbers.

With SAS 9.4 it should EM 12.3 and EGuide 6.  

See: SAS Enterprise Miner (product description).  and Model Development and Deployment, Data Mining Software, SAS® Enterprise Miner™ | SAS (see the requiments/bundling) and System Requirements--SAS® Enterprise Miner(TM) 12.3

You wil lsee the license of EM is including Eguide and AMO. The goal is that they cooperate.

More advanced: Text mining is requiring data mining (EM).

High performance procs are optional.  (VA not required).

Model manager is adding a level to it by managing the PMML-type models build with Eminer. 

See: Data Mining Group - Welcome to DMG for this dedicated area.

This is the SAS forum part of EM: https://communities.sas.com/community/support-communities/sas_data_mining_and_text_mining/

The difference between Eguide Eminer Amo is target group and the way users are expected to do their work.

Eminer is the ultimate tool for building the (PMML-style)  models following Semma/Crisp.

Eguide is the best tool for a developer/programer and analyst. Many times positioned as the switch army knife.

Amo  is the best for the casual SAS user mostly close to be an information consumer. Also available for SMB in a dedicated license   

[ChrisHemedinger]

Here's a blog post that restates what said:

A case of mistaken identity - The SAS Dummy

In recent years there has been even more integration: Rapid Predictive Modeler within EG and AMO, and with EG and AMO 6.1 there is now the Recency, Frequency, Monetary task (RFM), which is an entry-level customer segmentation model.  The RFM task does not require a license for SAS Enterprise Miner, so it's a great way to explore some possibilities without an additional investment.

Chris

[Elle]

http://support.sas.com/documentation/cdl/en/emgsj/67981/HTML/default/viewer.htm#n1cpd0rgpneqwqn16mfc...