Real Time Analytics - Basics for Updating Algorithms

From the point of implementing real time analytics, in a decision portal, we need to know how to update the decision algorithms based on data that are coming in real time. 

Requirements for Implementing Real Time Analytics:

The key principles that will facilitate application of real time analytics are the following:

  • Simple yet does not sacrifice the predictive power
  • The traditional algorithmic steps and associated formulae have to be re-configured and amenable for on-the-fly update of decision rules
    • The algorithms are different types; each one of them will require its own re-configured updating formulae; some of the key algorithms that are usable for real time decisions are
      • CART
      • Clustering Methods
      • Neural nets
      • Regression methods
      • Applications of Bayes Theorem and Its Extensions
  • The real time decision servers have to be different from the real time web-content server
  • No intervention from updating the algorithm to the implementation for the web delivery

Successful Consumer Interaction Requires Immediate CAR:

Web is the ultimate center of interaction and communications with consumers have to have three basic principles of communication

  • Consistent
  • Accurate
  • Relevant

(Communication involves driving C.A.R) for marketing to succeed.  The fourth important rule is that the interaction/communication be relevant right away. That is drive the CAR right away.

Examples of Applicable Real Time Analytics:

Some interesting examples that absolutely requires real time analytics are the following:

  • Delivering real time marketing content for a consumer
  • Delivering real time decisions to a patient over web
  • Delivering real time decision to security regarding the entrant to a building
  • Delivering real time decisions to the security as to the nature of evolving terrorist threat

For more information contact

Nethra Sambamoorthi ( or  )

11 Bartram Road

Englishtown, NJ 07726