Published Release Notes

New Text Analytics landing page

The Text Analytics landing page allows you to create the sentiment and classification text analysis models using a wizard. The wizard allows you to upload training data, train a model using different algorithms, analyze the accuracy of the model, and export the model. You can also create a binary rule file that contains the model and upload it as part of taxonomy or sentiment decision data.

For more information, see Text analytics functionality enhancements.

Free Text Model rule form enhancements

Enhancements to the Free Text Model rule in DSM expand the text analysis functionality. On the Select Analysis tab, you can include text analysis models in the sentiment or classification analysis. This option enables the combination of rule-based and machine-based learning approaches in text analysis and provides more accurate and reliable results. On the new Advanced tab, you can configure language detection settings, enable spell checking, set the score range for the neutral sentiment, and control the text categorization settings based on various criteria.

For more information, see Text analytics enhancements.

NLP outcome mapping enhancements

Enhancements to NLP outcome mapping in DSM allow you to better navigate to an exact entity or category occurrence and highlight overlapping sentiments. For sentiment analysis, the text is now highlighted in the color corresponding to the sentiment detected in a particular sentence or the overall sentiment of the document. For classification analysis, you can examine various elements identified in the analyzed corpus, like categories, entities, and so on.

For more information, see Text analytics functionality enhancements.

Strategy rule form Test Run panel enhancements

The enhanced Test run panel on the Strategy rule form enables an in-depth analysis of your strategy both in terms of the distribution of propositions among customers and strategy performance. For enhanced analysis of decisioning outcomes, you can now test your strategy on real data by selecting data flows and external input as the sources for the test run. For detailed performance analysis, you can use enhanced metrics with various statistics that provide insight into how the strategy execution impacts the performance of the system.

For more information, see Strategy rule enhancements.

Cannot view all custom context posts in Pulse

When you use Pulse in a custom context, the custom context label is displayed as plain text instead of as a link that you can click to view all posts that are associated with the custom context.

For more information about Pulse, see the help topic Pulse.

New Simulation Conversion wizard

The data flow runs replace simulations that are driven by the Interaction rule instances. Simulations now leverage data processing and strategy instructions that are defined in data flows. Use the Simulation Conversion wizard to convert legacy simulation objects into data flows and data sets. The converted legacy simulations can be re-executed using the batch processing option on the Data Flows landing page.

For more information, see Converting simulations.

New Decision Strategy Manager (DSM) alerts

The new DSM-related runtime alerts allow you to detect performance anomalies in your decisioning projects. The alerts are published in Autonomic Event Services (AES).

For more information, see Performance alerts, security alerts, and AES.

Improvements for the PMML models

The Pega 7 Platform extends support for the PMML models to include PMML 4.2 features and validations. Additionally, the PMML files that contain custom functions can be imported into the Pega 7 Platform and used in strategies.

For more information, see Predictive Model rule form.

Improvements to the adaptive models management

There are several improvements for managing large sets of adaptive models that make model management faster and easier. In the Adaptive Models Management landing page, you can clear and delete models in bulk. In the Predictors tab of the Adaptive Model rule, you can add multiple predictors. In the Memory estimation tab of the Adaptive Model rule, you can estimate the memory that the Adaptive Decision Manager (ADM) system needs to accommodate the models that use the adaptive model configuration.

For more information, see:

• Adaptive Models Management landing page
• Adaptive model form - Completing the Predictors tab
• Adaptive model form - Complating the Memory estimation tab

Improvements to the HBase data set, HFDS data set, and Hadoop host configuration

The Pega 7 Platform features maintenance improvements to the Hadoop host configuration, HBase data set, and Apache Hadoop File System (HDFS) data set. System architects can more easily make HDFS and HBase storage available to business scientists for exploratory analysis and predictive model building.

For more information, see Big data functionality enhancements.