First Look: GDS Link Decision Intelligence

April 24, 2013

in Analytics, Data Mining, Decision Management, Product News

Share

I got an update from GDS Link on their Decision Intelligence product recently. This was developed initially as a consulting tool and is now available as a licensed product. The purpose of Decision Intelligence is to help customers anticipate business corrections before there is an impact on the business. Decision Intelligence provides scorecard monitoring or what I generally call Decision Analysis. In fact the monitoring could be of either scorecards or complete decision services. These could be built with the GDS Link platform or another platform. The current version is a Java based application and is locally installed. It can be used to monitor scorecards at implementation and periodically over time rather than simply waiting and reacting when something goes wrong.

Decision Intelligence shares a data dictionary with Decision Studio (the GDS Rules Engine) and provides a graphical interface for the analyst that analyzes the outcome of the decision. The user can manage an unlimited number of decision services, scorecards and datasets and build an unlimited set of indicators for analysis of these scorecards and decision services. Several components are included:

  • An Analysis Editor where the user sets up the dimensions or variables, indicators and populations (so you can compare a consistent group over time). Any variable can be analyzed, not just the ones used in a scorecard, and all analyses are stored and tracked.
  • Performance Analysis for scorecard performance is designed to show overall performance. It compares populations (development v production populations for instance) and shows bad rates or other indicators. How these indicators vary by variables and value ranges is displayed and the results can be filtered, to show only accepted applications for instance. Charts are provided and standard measures such as GINI, Area Under Curve and K-S analysis are provided.
  • Stability analysis shows model and characteristic stability across populations (this year v last year, development v production etc). This also allows filtering by variables like date or accept status and is designed to find problems that result from shifting populations.
  • Characteristic analysis shows single variable performance, allowing analysis of performance by individual variables or characteristics. This allows an individual characteristic to be monitored to see if it has an unusual effect on a measure – to make sure that a higher value for the attribute always has a consistent effect on score for instance. Again standard measures like GINI, Area under curve and K-S can be graphed.
  • Finally there is a general analysis tools for cross-characteristic analysis using a cross-tab metaphor.

The whole environment is designed to be very graphical with a largely drag and drop development. As the tool is designed for ongoing decision analysis there is also easy development and management of a set of graphs/reports so that your preferred standard set can be run whenever you want. Data can be loaded in from multiple formats on a batch basis and tables and charts can be exported for easier reporting.

With its release as a commercial product, GDS Link Decision Intelligence will be in the next release of our Decision Management Systems Platform Technologies Report and you can get additional information on it here.

Share

Previous post:

Next post: