Today I am happy to announce the release of new versions of the main products in the IBM SPSS Data Science portfolio -- IBM SPSS Modeler 18.1, IBM SPSS Collaboration and Deployment Services 8.1, and IBM SPSS Analytic Server 3.1.
This new release presents six major categories of improvements:
Continuously embracing and extending open source capabilities - For the past few years IBM SPSS Modeler has extended and embraced open source capabilities - first with R and last year with Python and Spark. In this release we have continued commitment to bring the open source flexibility to the non-coding environment. We have introduced three Modeler nodes that run Python algorithms -- one-class SVM, SMOTE and XGBoost. These well-regarded algorithms, first only available via Python coding, are now exposed directly in the Modeler GUI which facilitates to the non coding audience to take advantage of the Python algorithms ease. Modeler 18.1 will ship with Python 2.7 so end users will not need to install it separately. We also have other important improvements with open source integration in this version, such as the ability to run Python code without creating an extension and the ability to run any version of R. Spark 2.0 is also incorporated into this release.
Text analytics for Big Data and Mac - In Modeler 18.1 we have migrated the back end of our text analytics engine to System T which is a multi-threaded more scalable engine. This will allow you scale to much larger datasets. We have also added a new node to detect the language of text -- and text analytics is now available for Mac users.
Improved Integration with The Weather Company Data and Decision Optimization - Decision optimization is an IBM offering that enables a user to turn the predictions/scores produced by Modeler into actual decisions. It does so by taking into account the outcomes and constraints our customers have in running their businesses. For instance Modeler can produce a score on whether a customer will respond to a marketing campaign. However, if there are ten proposed campaigns, it is not always a simple matter to assign each customer to a campaign. Some factors need to be considered to assign the customers to the right offer, such as the overall marketing budget, the cost for a particular offer, how many times a customer can be contacted, and others. Decision optimization allows you to optimize your results by executing several scenarios. In Modeler 18.1 we expose the ability to run OPL -- the code used in IBM CPLEX Optimization Studio -- directly into Modeler. The OPL code can be executed via Modeler and managed through IBM SPSS Collaboration and Deployment Services.
Another important integration improvement is with the Weather Company. Weather data can add lift to predictive models in many instances -- such as predictive maintenance, predicting insurance claims, predicting sales activity and others. In Modeler 18.1 we have added the ability to get both historical and forecast weather data from The Weather Company, which makes the access to weather data easier, faster and more complete.
Modeler ability to read in data from Hadoop. - Last year we declared support for reading data from BigInsights via BigSQL. Now Modeler will also support direct access to data in Hortonworks HDP via BigSQL, Hive and Cloudera Impala without the requiring SPSS Analytic Server. While you are now able to read Hadoop data sets with Modeler, SPSS Analytic Server does make your job more effortless with the ability to drive predictive modeling into the Hadoop cluster without coding and without data transfer.
Energize your analytics through usability improvements - As a former long time SPSS Modeler consultant myself, I have found there were certainly aspects of the software that caused time and effort to be spent not as productively as desired. We have prioritized certain key usability changes to ensure that a user's time is even more productive. An example is auto save/auto recover. In prior versions, if a user failed to save a stream before his or her computer crashed or was rebooted (for instance due to a patch install) all the changes were lost. With Modeler 18.1, streams are automatically saved in the background. If Modeler does not shut down properly, the user is prompted on whether or not to re-open the streams. Other usability improvements include partial transpose, a new ability to manage Modeler palettes, and enhancements when using Analytic Server.
Software currency-- We also have updated the operating systems and supported data sources in this release.
We hope you are excited with the new SPSS Modeler capabilities and that you are following us on this journey to make it the most powerful, flexible and easy to use Data Science platform for those that prefers a tool where coding is an option. Please stay tuned for more details about this release in our upcoming July Webinar.
For more information about SPSS Modeler contact: 1-800-543-2185, or email us at salesbox@us.ibm.com.
Click
here to try SPSS Modeler free for 30 days.
#Algorithms#AnalyticServer#datascience#Optimization#python#SPSSModeler#WatsonStudio