What's new in version 18.4.0?

IBM® SPSS® Modeler adds the following features and enhancements in this release.

  • Text Analytics flows created in Cloud Pak for Data can now be exported from Cloud Pak for Data (in JSON template format) and imported to a IBM SPSS Modeler stream.
  • You can now switch Python environments via the SPSS Modeler user interface. For details, see Switching to a different Python environment.
  • SPSS Modeler now provides enhanced password encryption. Due to a backward compatibility issue, note the following:
    • If you run SPSS Modeler Server using a private password file to authenticate users against a private password database, you must use the pwutil executable to recreate the database in 18.4. For instructions, see Configuring a private password database.
    • If you use an encrypted password (via the Encode Password option on the Tools menu), you must recreate it in version 18.4. For more information, see Generating an encoded password.
  • A new menu item is available for managing your SPSS Modeler Server connections: Tools > Manage Modeler Server Configuration. The available settings are similar to those found in the SPSS Modeler Server Configuration section of the Modeler Administration Console included with SPSS Collaboration and Deployment Services Deployment Manager. For details about the settings, see SPSS Modeler Server Configuration.
  • Empty passwords are no longer supported for Cognos TM1 Server connections.
  • Windows 11 is now supported.
  • MacOS Monterey (also known as MacOS 12) is now supported.
  • Java 11 is now supported.
  • CPLEX 22.1 is now supported.
  • InstallAnywhere 2021 sp2 is now supported.
  • SPSS Data Access Pack 8.1.1 is now supported
  • Netezza Performance Server 11.x is now supported.
  • Cognos Analytics Connector 11.1.7 is now supported.
  • Amazon S3 is now supported through Athena (for reading data, only).
  • ClickHouse database version 22.3 is now supported with the following limitations:
    • SQL pushback isn't supported for the Merge function in the database export node (due to weak support by ClickHouse for update, delete, etc.)
    • SQL pushback isn't supported for the time type (due to there being no time type in ClickHouse)
    • The date baseline is 1970-01-01