- Teradata Query Banding in Database connection Presets.
- Updated GLMM node options.
- Support for encrypted and compressed .sav files.
- Support for SPSS Modeler Server single sign-on (SSO) without requiring IBM SPSS Collaboration and Deployment Services.
- Database layer support for Single Sign-On.
- SQL Pushback for the Sample node for zDB2.
- R nodes are now installed as part of the base SPSS Modeler installation.
- Modeler Adapters for IBM SPSS Collaboration and Deployment Services are installed via IBM Installation Manager.
- Support for Russian localization.
- Option to enable Federal Information Processing Standard (FIPS) encryption. Secure Sockets Layer (SSL) configuration to SPSS Modeler Server updated to use Global Security Kit (GSKit); replacing OpenSSL.
- Support for IBM SPSS Collaboration and Deployment Services Context Root feature – the connection dialog box now requires a URL instead of specifying a hostname and port number.
- IPv6 support added
New features in IBM SPSS Modeler 16
Monte Carlo Simulation. A new Simulation Source node provides an easy way to generate synthetic data from scratch using a wide selection of statistical distributions. Alternatively, the new Fitting node can automatically build a preconfigured source node reflecting the distributions of and relationships between historical variables. The Simulation Evaluation node is a terminal node designed to evaluate fields resulting from a simulated analysis stream, and provides useful distribution and correlation charts. Python Scripting. Scripting in IBM SPSS Modeler, used for automating processes in the user interface, can now use the Python language, as well as continuing to support the legacy scripting mode. Python is a well known and popular language that provides a rich set of features including a rich and concise syntax, error handling, and powerful add-on modules. Note: The Script tab in Tools > Stream Properties is now named Execution. Looping and Conditional Execution. These new options enable simple looping and conditional execution of streams without needing to code scripts. Find these new options in Tools > Stream Properties > Execution or by right clicking a node within a stream and choosing the Looping/Conditionalexecution option. Space-Time-Boxes (STBs) node. Create bins of location and timestamp data to support more sophisticated analyses. In hangout mode, this node also identifies times and places where entities dwell. Additional expression builder functions support the extraction of STB centroids as well as geohashing. Entity Analytics Enhancements. As well as resolving individual entities this can now identify n-degree relationships between entities. Additional support is provided for real time updating via the streaming node, flattening resolved entities (Distinct node), and for anonymizing data as it is fed into an entity repository. Note: the previous local SolidDB database has been replaced with DB2. New Receiver Operating Characteristic (ROC) Evaluation node chart type and Area Under the Curve (AUC) and Gini metrics in Analysis node. Supports binary targets. New Distinct node option for creating a composite record. Enables you to specify the a method of aggregation for each field being grouped (first value, last value, concatenate values, and so on) rather than discarding duplicate records. Whereas the Aggregate node is typically used for summarizing data to a higher level, this new option is used for flattening duplicates; for example, those identified through entity resolution. TM1 Source and Export nodes. Enables you to access to TM1 Cube Views via the TM1 Source node, and score data back to an existing TM1 cube using the TM1 Export node. Aggregate expressions and Window Aggregate functions. You can create custom aggregation expressions in the Aggregate node, incorporating built-in aggregate functions (MEAN, SUM, and so on) and/or Database Aggregate User Defined Functions. In derived expressions you can derive fields that require windowed aggregation functions (such as moving averages). Built-in and database-provided window aggregate functions are available. IBM Netezza Analytics in-database mining enhancements. New Netezza Two Step algorithm, additional Helper Application option to manage (delete, rename, and so on) Netezza analytic models, support for Model viewers for Regression trees, Decision Trees, Kmeans, and TwoStep. Note: New features require INZA 3.0. New R nodes and Custom Dialog Builder for R. In addition to the R Model Build node and model nugget introduced in SPSS Modeler 15 Fix Pack 2, this release adds 2 new nodes – R Process and R Output. With the R Process node, you can take data from an SPSS Modeler stream and apply transformations to the data using R scripting. With the R Output node, you can use your own custom R scripts to perform data analysis and to summarize the results of model scoring. You can produce text and graphical output of your analyses. This output can be directed to a file, or viewed in the R Output Node Output Browser. The Custom Dialog Builder provides the ability to create custom Model Build, Process, and Output node types and model nuggets, including field chooser, text, number, radio buttons, and sub dialog controls to enable abstraction and parameterization of R programs. You can choose the node type, destination palette, and node icon before installing the node or sharing the node for use by other SPSS Modeler users. The custom dialog builder is launched from the Tools menu. Note: To use this feature, you must have installed SPSS Modeler – Essentials for R. R in database. SQL pushback support for R nodes; for Netezza, SAP Hana and Oracle by utilizing their R support. Note: Databases need to have the appropriate vendor-provided R extensions installed. Streaming Time Series Process node. Build and score time series models in a single step to provide real time deployment through IBM InfoSphere Streams, the IBM SPSS Collaboration and Deployment Services Scoring Service, or IBM SPSS Modeler Solution Publisher. Preview button. When used in conjunction with a database source the preview enables SQL pushback to be visualized. When used in conjunction with an Analytic Server data source, use the Preview button to receive information on potential large data movements. New Analytic Server options in the Auto Classifier, Auto Numeric, and Auto Cluster nodes. When running a stream against IBM SPSS Analytic Server, you can choose between running with Splits enabled (if you are using the split model feature, use this option ) or Very Large Data options (splits are ignored, and the modeling objectives are set for Big Data). Scoring is now supported for Auto Classifier, Auto Numeric, and Auto Cluster models that can be built in SPSS Modeler Server. Enhanced Scoring Adapter support. Support for scoring Text Mining and imported PMML models via Database Scoring Adapters. Also new Database Scoring Adapters for DB2 LUW. Data View source node. The Data View node enables you to read data from an Analytic Data View into an SPSS Modeler stream. The Analytic Data View is a new way to create a unified data view in IBM SPSS Collaboration and Deployment Services 6. This feature replaces the Enterprise View node, which is no longer displayed on the node palette, but is supported in streams imported from previous versions of SPSS Modeler. Miscellaneous other enhancements.