Watson Studio, Watson ML, Watson OpenScale

 View Only

Announcing IBM SPSS Modeler 18

By Ted Fischer posted Tue March 15, 2016 05:05 PM

  
Today we are releasing Modeler version 18. There a quite a number of important changes and improvements in this version. We have four groupings of changes – Big Data Algorithms in Modeler, changes that continue Extend and Embrace the Value of Open Source, Platform Flexibility and other changes.

Big Data Algorithms in Modeler

Over the past a year, a number of algorithms were added to Modeler but with the restriction that they only run with Analytic Server –which is the connector from Modeler to Hadoop. In version 18, all six of these algorithms are now available in Modeler with any type of data. The algorithms include

• Random Trees – a popular method in the data science community that involves taking a C&R Tree model with bagging and then only consider a sampling with replacement of variables for each split of the tree
• Tree-AS which is based on CHAID
• GLE – which incorporates a number of regression methods
• Linear-AS which performs linear regression
• Linear Support Vector Machines
• Two-Step-AS clustering

An important feature of all these algorithms is that they are multi-threaded – i.e. a single build can use more than one core. This will improve model build times for large data sets and make better usage of data resources. GLE and Linear SVM support regularization which prevents overfitting by penalizing models with extreme parameter values. Finally, Tree-AS and Linear SVM have behind the scenes data preparation that will automatically handle common data issues

We have also added a big data algorithm in Modeler version 18 not present in version 17.1– a new version of the time series algorithm. Like the old version, it supports three methods of forecasting exponential smoothing, ARIMA and expert Modeler. In version 18, time series will run in Analytic Server and support multi-threading. In addition, the new algorithm supports split modeling. In Modeler, a variable can be defined as a split variable in the type node – with the result that supported algorithms will then produce a separate model for each split. With version 18, time series can be added to this list of supported algorithms.

time_series_split

Extend and Embrace the Value of Open Source


For many years we have been extending and embracing the value of open source. As you can see in this community, we have many open source extensions that allow non-programmers to run open source programs to do anything from modeling to different graphs to getting different types of data. We started extension in version 16 with R extensions. In version 17.1, we added Python with Spark extensions but required them to run in Analytic Server. Now with version 18, Python with Spark extensions will run natively in Modeler. We have also included Spark within the Modeler download so that any Python code can access Spark machine learning libraries – note that a Python 2.x must be installed separately. The distribution that we have used in testing is Anaconda found at https://www.continuum.io/downloads.

With this change, all Modeler users can now run Python extensions. They can invoke the Spark machine learning libraries that include many algorithms not found in Modeler such as gradient boosted trees. If the appropriate Python libraries are installed, data scientists can also invoke common Python machine learning libraries such as num-py, scipy, scikit-learn and Pandas.

Python_custom_node

We have also made it easier now to get extensions from the community. Using the new Extensions menu item, Modeler users can now invoke an Extension hub. With this hub, users can identify, download and install extensions without having to go to Github and manually transfer file.

extension_hub
Platform Flexibility

We have added a couple of links in the Help menu to this community – particularly to the forums and the community help page.

links_to_community

Modeler Personal and Professional will be available on Mac OS with version 18. In addition, all versions of Modeler 18 support Windows 10.

mac_version_18

Other Changes

Modeler 18 extends its in-database mining capabilities to include DB2 in Z/os or IDAA (IBM SB2 Analytics Accelerator). Using a GUI, Modeler customers can now build and deploy models using the Decision Tree, Regression Tree, K-Means, Native Bayes, and Two-step algorithms.

Modeler Premium now includes additional entity analytics capabilities – including the ability to use an external DB2 repository, more than 4 cores and exposing relationships. Please note though that usage for more than 10 million records is no longer recommended.







#Algorithms
#Programmability
#python
#Spark
#SPSSModeler
#WatsonStudio
45 comments
13 views

Permalink

Comments

Tue May 22, 2018 08:30 PM

This should work -- but you should definitely test the streams before deploying them.

Tue May 15, 2018 09:03 AM

Can legacy version 14.2 streams and scripts be run in version 18?

Sun February 11, 2018 07:35 PM

Please go to Passport Advantage to obtain your copy.

Sat February 10, 2018 07:03 PM

Where can I download the IBMS SPSS Modeler v18? Is Passport the only means to get my own copy

Tue May 23, 2017 10:32 AM

When is the Modeler version 19 expected release date?

Tue May 23, 2017 09:04 AM

How can we integrate Python pandas in SPSS Modeler v18??

Fri January 06, 2017 06:02 AM

AS 3.0 supported HDP 2.3.2 and Ambari 2.1.1 when eGAed on 3/15/2016. The Ambari version always goes with the Hadoop distribution. At that time HDP 2.3.2 was supporting Ambari 2.1.1. At AS 3.0 refresh on 6/14/2016, HDP 2.4 also supported.

Thu January 05, 2017 09:18 AM

No, not at this time, although it is a known enhancement for a future release. For now, the workaround is to use a dummy user input node in front of the process node.

Thu January 05, 2017 09:17 AM

Yes.

Thu January 05, 2017 06:45 AM

Does the streams developed under SPSS Modeler 16.0 are fully compatible with SPSS Modeler 18.0 ?

Sat July 09, 2016 05:52 AM

Can source nodes be created with R?

Thu May 26, 2016 02:28 PM

Does IBM SPSS Analytic Server 3.0 support
Hortonworks HDP 2.3.2 version and Apache Ambari version 2.2.2.?
Have Modeler v18 installed, now want help to proceed with AS.

Fri May 06, 2016 03:12 AM

Useful comments ... Thanks Ted Fischer ..!

Sat April 30, 2016 02:02 AM

The jython implementation in Modeler v17.x was based upon python v2.6, if I recall correctly. Which python version is jython based upon for Modeler v18.x?

Thu April 28, 2016 05:07 PM

Only Python 2.x is supported with Python in stream.

Thu April 28, 2016 04:06 PM

Installed Modeler v18 with anaconda python v3.5. Then modified the options.cfg to refer to python.exe.
Post this used example of k-means(got an eas.cpp error) and with the CF example (got some AS api error). Examples downloaded from the extension site.
Want to execute a sample to use numpy and then MLlib on just the Modeler client. Hope I got it right that without Analytics server we could execute these examples.

Thu April 28, 2016 02:55 PM

The actual installation process for Modeler client is fairly similar -- simply pick the concurrent licensing option in the installation routine. However, using a concurrent license in Modeler does require a license manager to be installed and activated first.

Thu April 28, 2016 02:53 PM

Please see this blog entry for some more detail on coding a Python extension

https://developer.ibm.com/predictiveanalytics/2016/03/07/coding-a-pythonspark-modeler-extension-for-collaborative-filtering/

Thu April 28, 2016 02:51 PM

Yes

Thu April 28, 2016 11:17 AM

Is the installation of SPSS Modeler 18 with single user license is different from concurrent users ..??

Fri April 22, 2016 12:42 PM

In general, you can look for prerequisites through the compatibility reports at http://www-969.ibm.com/software/reports/compatibility/clarity/index.jsp. For C&DS 8, the minimum version of DB2 is 9.7 and the minimum version of WAS is 8.5.0.2

Fri April 22, 2016 05:55 AM

Thanks Ted for the reply.

I guess we need to update the IBM CaDS to v8, for deploying the streams built using v18.0.
Also what is the supported minimal IBM WAS and DB2 version with CaDS 8.0.

Thu April 21, 2016 09:08 PM

You can connect a Professional client to a Premium Server. However, the Professional client will not be able to access the Premium features.

Wed April 20, 2016 05:14 PM

Is it possible to connect from any SPSS Modeler client (either Professional or Premium) to SPSS Modeler Server Premium ? We are planning to have the Server versions with Premium and restrict the clients to use Professional or Premium based on use case scenario.

Tue April 19, 2016 03:08 PM

@ZoHa: I am playing with this new functionality, and found the following. As for a sample stream: you can download the collaborative filtering extension, including an example stream, from the Extension Hub. The script is described in https://developer.ibm.com/predictiveanalytics/2016/03/07/coding-a-pythonspark-modeler-extension-for-collaborative-filtering/. (The article is about IBM SPSS Modeler 17.1, but in 18 you can run it against IBM SPSS Modeler Client - provided you have Python 2 (e.g. Anaconda) installed, and provided you have configured IBM SPSSS Modeler Client to use your Python installation (refer to page 12 of the ModelerExtensions.pdf file, that ships with the software)).
@Mehrdad. Yes, and furthermore (from the ModelerExtensions.pdf file, page 11): "If you want to use the Machine Learning Library (MLlib), you must install a version of Python that includes NumPy."

Hope this helps, Jos

Tue April 19, 2016 06:43 AM

Can we import GraphX using SPSS Spark ?

Thu April 14, 2016 05:46 AM

Can We use Python's librarys such as Numpy, Scipy and etc?

Wed April 13, 2016 06:15 PM

"What is new here is n ability to run Python in stream and that was tested with Anaconda." - can you provide a tutorial for this or a sample script?

Wed April 13, 2016 04:55 PM

SPSS Modeler Premium 18.0 (Install Base Modeler) - I see only this in the standard software installer.

Sorry for being so persistent, I really need to have the text mining node.

Wed April 13, 2016 04:37 PM

Can you provide a sample stream or tutorial about how to run Python script from Modeler Client? Does it work on Modeler Server too?

Wed April 13, 2016 03:42 PM

For Modeler you can do that.

Wed April 13, 2016 03:42 PM

The existing Python installation (or rather Jython) for scripting is still present. What is new here is n ability to run Python in stream and that was tested with Anaconda.

Wed April 13, 2016 01:17 PM

Soundarya, The installation is performed in two steps. First install the Modeler Client - the download is about 1GB. Next, install the Premium Client - this is about 340MB. The second installation will give you Text, EA, and SNA.

Mon April 11, 2016 02:12 PM

Hi Jos, this is available in the client

Mon April 11, 2016 02:08 PM

When you say that you've tested with Python 2.x from Anaconda did you replace the existing python distribution the the new one ?

Mon April 11, 2016 12:00 PM

Hi, significant :-) improvements, great job! Just to make sure ... all this new functionality is available in the Client version of IBM SPSS Modeler, is that correct? Or is IBM SPSS Modeler Server required at some point? Thanks, Jos

Mon April 11, 2016 05:52 AM

Hi, I was able to install modeler 18, but I dont see the text mining node. I had the text mining node installed for 17.1. Any help on this?

Sun April 10, 2016 02:53 PM

Its pretty awesome that SPSS Modeler now runs on Mac OSX. Thats important news!!

Thu April 07, 2016 12:41 PM

yes

Thu April 07, 2016 10:47 AM

Can we upgrade from version 16 fp2 to 18.0 directly, without moving onto v17.X?

Tue April 05, 2016 04:37 PM

Yes

Tue April 05, 2016 02:57 PM

Hi,

Does the streams developed under SPSS Modeler 17.0 are fully compatible with SPSS Modeler 18.0 ?

Wed March 30, 2016 07:19 AM

Hi Steve, thank for your reply.

I have a windows OS TP only. But dont see SPSS modeler 18 listed in ISSI.

Sat March 26, 2016 02:44 PM

You can use ISSI but only in a Windows OS

Wed March 16, 2016 09:29 AM

Hi,
Can I install this from the standard software installer in IBM?