Monitoring for event patterns


Having employed the right analysis algorithms and methods for identifying a one of a number of interesting event patterns, the next step is enabling an ability for recognizing those patterns. This would be pretty straightforward if each event pattern was discrete and always took place in a defined total order, but this is not frequently the case. If you recall our banking attrition example, there were a number of banking transactions that could precede the severance of the customer’s relationship – an address change, account consolidation, reduction or elimination of online transactions. But they don’t necessarily happen in one specific order.

What is needed is a framework that can integrate these key capabilities (among others) associated with maintaining the status in relation to transacted events:

  • Simultaneous monitoring of many concurrent activities
  • Embedded states representing stages in different sequences of events, even if they do not take place in a specific order
  • Integrated business rules to be applied once a particular state is reached
  • The ability to ratchet back into a previous state once some conditions are met (such as an expired time frame)

Fortunately, there are methods and systems that can accommodate these requirements. An event stream processing (ESP) system can incorporate the representations of a pattern sequences of events and transactions as a collection of phases in a giant state transition diagram, allowing state transitions whenever one event of a defined pattern takes place. A particular state will be reached when all of the events in the pattern have been identified, and a business rule can execute to notify a customer support representative with the specific details along with some suggestions for remedial actions that can be taken.

To use the same banking attrition example, a change of address and account consolidation may be indicative of a customer who is moving out of the bank’s branch service areas. If the customer is moving to a different location, he/she may have decided to close the accounts and open a new account at the new location. However, if that same customer has also been a frequent user of online banking, perhaps there are opportunities to create an account relationship that preserves the online banking relationship, such as offering no-fee use of other ATM networks and fee-free cross-bank money transfers.

Coupling market-basket and event stream analytics with event stream processing provides a sound approach to predictive modeling that can improve numerous business processes. While both rely on very different fundamental algorithms, a savvy business manager will combine their use to help integrate the anticipation of particular outcomes into improved business processes.


About Author

David Loshin

President, Knowledge Integrity, Inc.

David Loshin, president of Knowledge Integrity, Inc., is a recognized thought leader and expert consultant in the areas of data quality, master data management and business intelligence. David is a prolific author regarding data management best practices, via the expert channel at and numerous books, white papers, and web seminars on a variety of data management best practices. His book, Business Intelligence: The Savvy Manager’s Guide (June 2003) has been hailed as a resource allowing readers to “gain an understanding of business intelligence, business management disciplines, data warehousing and how all of the pieces work together.” His book, Master Data Management, has been endorsed by data management industry leaders, and his valuable MDM insights can be reviewed at . David is also the author of The Practitioner’s Guide to Data Quality Improvement. He can be reached at

Leave A Reply

Back to Top