Angoss Software Corporation has announced the release of version 9.5 Platform, which includes enhancements to current KnowledgeSEEKER™, KnowledgeSTUDIO™, and KnowledgeREADER™ products, as well as the launch of two newly developed products, InsightOPTIMIZER™ and KnowledgeMANAGER™. Version 9.5 product and feature additions will considerably augment the current Angoss Data Science Platform offering, and continue to improve the user experience.
Further extending Angoss’ best-in-class segmentation and modeling interface to improve performance and increase business productivity, KnowledgeSEEKER, KnowledgeSTUDIO and KnowledgeREADER will now incorporate Ensemble Trees and redesigned Weight of Evidence (WOE) transformations that substantially enhance the automation of the scorecard building process. Additionally, in order to cater to a wider audience and increase collaboration between teams and departments, these products will also posses optional Python language integration capability.
In the new 9.5 Platform, Angoss extends its text analysis product, KnowledgeREADER, by providing text analytics users with new technology offered by the upgraded Lexalytics’ Salience 6 Text Analytics Engine. Compared to Salience 5 featured in the 8.7 Platform, the Salience 6 Engine in KnowledgeREADER 9.5 supports new features such as extended multilingual support and enhanced Syntax Matrix technology that enables users to quickly and efficiently extract meaningful and actionable results from text. Furthermore, text analytics features are now fully supported by the Angoss workflow. The new Text Analysis palette includes the Text Analysis, Association Map, Text Analysis Comparison, and Trend Analysis nodes.
Angoss is excited to unveil InsightOPTIMIZER and KnowledgeMANAGER, new product additions to the current software suite portfolio that offer highly anticipated optimization and model management capabilities while still being easy to use and accessible to technical and business users alike.
InsightOPTIMIZER, a prescriptive analytic application, optimizes the business strategy development process. It will help users solve complex business problems by automating the decision-making process, determining the next strategic move, and eliminating the need to experiment with different deployment scenarios thereby minimizing loss and maximizing results. By optimizing sophisticated analytic models, subject to business constraints and rules, InsightOPTIMIZER allows companies to make the best use of limited resources.
KnowledgeMANAGER will enable users to store and easily access their Angoss models via a user-friendly web-based interface. This application will allow users to manage, compare, monitor and score models to ensure that only the best performing models move to production. To adhere to heightened regulatory requirements, KnowledgeMANAGER provides managers with access to streamlined reporting and comprehensive audit trails to reduce compliance risk and deliver smarter decisions.
“The addition of the above new products and feature enhancements to our product portfolio has strengthened our competitive positioning in the Predictive and Prescriptive Analytics market with the offering of a complete Data Science Platform.” said Jim McGill, President and Chief Executive Officer, Angoss. “Our aim is to design solutions for both the technical and business users and truly provide products that are easy-to-use and deliver faster results than the competition.”
Features and improved capabilities in 9.5 Platform include:
KnowledgeSEEKER™ and KnowledgeSTUDIO™
Ensemble Trees
Random Forest, Bagging, and Boosting algorithms building ensemble decision trees
Python Language Integration
Decision Tree Enhancements
Force Split: Optimal Binning option
Option to allow or prevent split on the same variable on two consecutive levels
Tag Node and Node Comment feature improvements
Workflow Management Enhancements
Node alignment
Show/hide grid; Snap nodes to grid in the workflow canvas
Copy workflow layout to the same or other workflows
Cut and paste nodes; Copy datasets from other projects
KnowledgeSTUDIO™
New stopping criterion for Stepwise Regression: Entropy difference
Enhanced Scorecard Building Capabilities
System Stability Report
Improved and redesigned Weight of Evidence transformations
A separate node for WOE transformation
Option for manual editing of WOE values
More charts and statistics available in the Weight of Evidence Editor
WOE Transformations Report
Enhanced WOE Optimization features
Dynamic WOE binning adjustments in regression models
Scorecard scaling option: Incorporate Base Score into score points (introduced in 9.4.x)
KnowledgeREADER™
Text Analysis supported in Workflow
The Text Analysis palette includes the Text Analysis, Association Map, Text Analysis Comparison, and Trend Analysis nodes
Lexalytics’ Salience Text Analysis Engine upgraded to Version 6
Enhanced Syntax Matrix technology
New language packs – Italian, Dutch, Japanese, Korean, Malay and Singlish languages in addition to previously supported English, German, French, Spanish, Portuguese and Chinese.
In addition to all these NEW and enhanced features with Angoss Software Suite Version 9.5, Angoss is thrilled to announce the launch of two NEW products; KnowledgeMANAGER™and InsightOPTIMIZER™
KnowledgeMANAGER™
Manage Models – Enables users to oversee multiple models, as well as add new models, create champion – challenger models and close models
Compare Models – Compare competing models and select the best performing model for deployment
Monitor Models – Visual monitoring and alerts enable users to measure the accuracy and performance of models
Score Models – Automate scoring in batch or real time for immediate model outputs
InsightOPTIMIZER™
InsightOPTIMIZER handles linear and nonlinear user-defined functions with any constraints. The range of optimization problems covered are those on the level of aggregating the values over a given datasets or on the level of each record in a dataset.
For linear optimization, Angoss uses a linear programming solver based on the revised Simplex method. Nonlinear optimization uses the Boundary and linear equality/inequality constrained optimization.