What's new and changed in SPSS Modeler?
SPSS Modeler updates can include new features, bug fixes, and security updates. Updates are listed in reverse chronological order so that the latest release is at the top.
You can see a list of the new features for the platform and all of the services at What's new in IBM Cloud Pak for Data?
Initial release of Cloud Pak for Data Version 4.0
A new version of SPSS Modeler was released as part of Cloud Pak for Data Version 4.0.0.
Assembly version: 4.0.0
This release includes the following changes:
- New features
Version 4.0.0 of the SPSS® Modeler service includes the following features and updates:
- Environment size
- The default option has been changed from 4 vCPU + 12gb of RAM to 2 vCPU + 8gb of RAM. If you have compute intensive workloads that require more vCPU—such as Auto Classifier nodes, Auto Cluster nodes, Auto Numeric nodes, Auto Data Prep nodes, or Text Analytics nodes—we recommend you define a custom environment size.
- Interactive decision trees
- An interactive tree builder is now available for the C&R Tree, CHAID, and QUEST nodes. These
nodes use decision tree models to develop classification systems that predict or classify future
observations based on a set of decision rules.
Previously, you could only generate a tree model automatically, where the algorithm decides the best split at each level. Now you can use the interactive tree builder to take control, applying your business knowledge to refine or simplify the tree before saving the model nugget.
For details, see The interactive tree builder.
- New sample projects and tutorials
- The SPSS Modeler documentation includes tutorials that are based on sample projects that you can download. For details, see the SPSS Modeler tutorials.
- Simulation Evaluation (Sim Eval) node
- The Sim Eval node is now available for evaluating continuous fields. For more information, see Sim Eval node.
- Streaming TCM node
- The Streaming TCM node has been added for building and scoring temporal causal models in one step. For more information, see Streaming TCM node.
- Support for external R and Python libraries
- You can now load R and Python libraries to use with the extension nodes. Also note that R 4.0 is now supported. For more information, including instructions for installing packages that your scripts require, see Extension nodes.
- Upload streams
- If you use SPSS Modeler desktop Version 18.3, you can now upload streams to Cloud Pak for Data directly from the desktop user interface. For details, see Saving streams to Cloud Pak for Data.
- Usability improvements
- The following interface changes have been made to improve usability:
- Copy and paste and undo/redo functionality
- Improvements to the Expression Builder interface
- Analysis node output improvements
- Output preview improvements
- Building custom images to install ODBC drivers
- You can now build custom images based on the SPSS Modeler runtime images available in IBM Watson Studio. You can use custom images to install custom ODBC drivers. To create a custom image, you need to download the image of the SPSS Modeler runtime that you want to customize, build a new custom image by adding ODBC drivers to the image you downloaded, register the new image, and finally update the environment definition you created in your project to use the new custom image. For details, see Building custom images.
- New setting in the GLE node
- A new setting called Perform non negative least squares is now included with GLE node in the build options under Parameter Estimation. Non-negative least squares (NNLS) is a type of constrained least squares problem where the coefficients are not allowed to become negative. Not all data sets are suitable for NNLS, because NNLS requires a positive or no correlation between predictors and target. For more information about using the GLE node, see GLE node.
- Continuous machine learning
- Model drift is the process by which models become outdated as your data changes
over time. SPSS Modeler provides
continuous automated machine learning to help overcome model drift.
A result of IBM research, and inspired by natural selection in biology, continuous machine learning is now available for the Auto Classifier node and the Auto Numeric node.
For details, see Continuous machine learning.