Introduction

Firms that handle large quantities of data want to analyze data for specific behaviors. The Scenario Manager enables a business to define how these behaviors manifest themselves in their data and to detect possible occurrences of these behaviors.

Elements Used in the Detection Process

Our software uses sophisticated pattern recognition techniques to identify behaviors of interest, or scenarios, which are indicative of potentially interesting behavior. These behaviors can take multiple representations in a firm’s data. A pattern is a specific set of detection logic and match generation criteria for a particular type of behavior. Using the Scenario Manager, you can create one or more patterns to implement a scenario. Patterns access data by datasets, which are views that are defined in the Scenario Manager.

The application generates an alert to package one or more matches for analysis and action in the Web Application. If Oracle Financial Services discovers multiple matches for the same focus, processing combines the matches to create a single alert and facilitate the analysis of the found behavior.

To facilitate analysis of alerts, the application groups scenarios that represent related business problems into scenario classes.

Defining the Detection Tools

The Scenario Manager includes tools that you can use to perform data detection. Business firms use these tools to locate and analyze behaviors of interest to detect unusual activity within their business. These tools include the following:

• DataSet Editor: Select data sources, filter data, add derived attributes, and join data sources to create virtual tables of data. This data enables you to search for and analyze behaviors of interest.
• Sequence Scenario Editor: Define a behavior that consists of events in a predetermined order. You can use these events to thread multiple data streams together to find behaviors of interest.
• Rule Scenario Editor: Access pattern elements that you need to create when using this editor for detecting behaviors of interest.
• Link Analysis: Use the data mining technique to find networks of entities related in an interesting or unusual manner.
• Job Editors: Define a collection of instructions for executing the scenarios defined in the tools to detect and generate alerts.
• Threshold Manager: Add a threshold or threshold set to a pattern or dataset to define tunable values. A threshold provides flexibility for changing variable values. Changes to values can be made by the Threshold Editor in the Administration Tools without having to change the values defined in the Scenario Manager at the dataset or pattern level.
• Augmentation: Add additional data to existing matches (such as account number or the source of a deposit) so that a firm can analyze their matches in familiar terms that are understandable (that is, adding the name of a trader rather than displaying the TRADER_ID attribute in the dis- play).