20 February 2016

Predictive Intelligence: Data or Precogs...

The use of the term "Predictive Intelligence" has been around for a few years in the Operational Risk Management (ORM) community.  Born from the marketing collateral of the Business Intel (BI) vendors, it essentially requires hundreds of gigabytes or even terabytes of historical data and then is analyzed or data mined for so called insight.  The question is, why is this "Predictive Intelligence" and not just more "Information" in a different context?

Now introduce the nexus of our own "Trust Decisions" and the "Human Factors" associated with the science of cognitive decision making.  How do we as humans make our decisions to trust vs. how computers make their decisions to trust?  Are they not executing rules written by humans?  When is it information in a different format as opposed to true intelligence?

Christian Bonilla may be on to something here:
"Professionals in the foreign intelligence community take pains to distinguish between information and bona fide intelligence. Any piece of knowledge, no matter how trivial or irrelevant, is information. Intelligence, by contrast, is the subset of information valued for its relevance rather than simply its level of detail. That distinction is often lost in sector of the enterprise technology industry that is somewhat loosely referred to as Business Intelligence, or BI. This has become a bit of a catchall term for many different software applications and platforms that have widely different intended uses. I would argue that many BI tools that aggregate and organize a company’s information, such as transaction history or customer lists, more often provide information than intelligence. The lexicon is what it is, but calling something “intelligence” does not give it any more value. In order to sustainably outperform the competition, a company needs more than a meticulously organized and well-structured view of its history. Decision makers at all levels need a boost when making decisions amidst uncertainty and where many variables are exerting influence. They need what I would call predictive intelligence, or PI – the ability to narrow down the relevant variables for analysis and accurately measure their impact on the probability of a range of outcomes."
What does the fusion of human factors have to do with predictive intelligence?  That depends on how much you value the kind of innuendo and messages in the Tom Cruise movie, Minority Report.  Many aspects of the original Philip K. Dick story were adapted in its transition to film that was filmed in Washington, DC and Northern Virginia.  Is it possible to predict someone's future behavior even before they commit a crime or even become violent?
Set in the year 2054, where "Precrime", a specialized police department, apprehends criminals based on foreknowledge provided by three psychics called "precogs".
Cruise plays the role of John Anderton who is part of the experimental police force known as "Precrime."  These aspects of clairvoyance and precognition has many skeptics and their use for predicting future events or a related term, presentiment, refers to information about future events which is said to be perceived as emotions.
Regardless of terms, beliefs or whether the software analytics are using historical data, the science of "Predictive Intelligence" is about forecasting the future.  Based upon the recent global events that missed the forecast of economic implosion based upon historical data, maybe it's time to start introducing more human factors to the equation.

The interviews with people who have gone on record to predict a future historical event will probably be right at some point in time. How long will you be around to wait?  The demise of the banking sector and the extinction of Lehman Brothers, Bear Stearns and maybe even AIG were most likely predicted by someone, somewhere in 2007/2008 time frame.  The point is that you have to have context and relevance to the problem being solved or the question being asked.
The real story of the crash began in bizarre feeder markets where the sun doesn't shine and the SEC doesn't dare, or bother, to tread: the bond and real estate derivative markets where geeks invent impenetrable securities to profit from the misery of lower--and middle--class Americans who can't pay their debts. The smart people who understood what was or might be happening were paralyzed by hope and fear; in any case, they weren't talking.
Predictive analytics extracts relevant information from data and attempts to forecast the future. It relies on capturing relationships between explanatory variables and the predicted variables from past occurrences, and exploiting it to predict future outcomes.  Is it possible that there was and is too much reliance on the numbers and not enough on people's cognitive intuition?

This blog has documented the "11 Elements of Prediction" in the past.  Now it's time to utilize the combination of these human factors in close collaboration with the data analytics and raw numbers. Effective execution of both will provide corporate management the situational awareness they seek within the time line they wish.

The future state of Predictive Intelligence will combine the science of "Trust Decisions" with the art of "Data Analytics" to achieve our desired outcomes.

No comments:

Post a Comment