In Praise of Bayes

Sr. Product Marketing Director


Statistical techniques such as predictive analytics help us make sense of ever-growing data. Analyzing signals – data streams created as events unfold –in-context of long historical records create insight to likely outcomes, and guide our decision-making. Let’s take a minute away from the hubbub of the here-and-now to reflect how this is made possible by a creative mathematician working in the 18th Century.

On this day, April 17, 250 years ago the Reverend Thomas Bayes dies in Tunbridge Wells, England. A Nonconformist, Thomas practices outside religious observances prescribed by Parliment for the Established Church of England in the 1662 Act of Uniformity. This, I think hints at the character of the man. He participates in the Enlightenment, creating a discourse spanning Europe and across the Atlantic to the American colonies. Exchanging ideas, establishing new values, these men and women of the mid-eighteenth Century create a public sphere of dialog, beyond control of Crown, State and religious institutions, wherein culture and dogma prove too heavy mulch for germination of new ideas. Thomas publishes two books, one on theology, the second, a defense of Isaac Newton’s calculus in response to criticism by Bishop Berkeley. In later life Thomas gives much thought to the mathematics of probability, but doesn’t publish.

Inheriting Thomas’ manuscript, his friend Richard Price discovers his work on inverse probability including a number of special cases Bayes uses to deduce his theorem. Price, a moral philosopher, influences John Adams, Benjamin Franklin, Thomas Jefferson, Thomas Paine, William Pitt, and Mary Wollstonecraft amongst many others. Price edits Bayes’ work and in 1763 reads his “Essay towards solving a problem in the doctrine of chances” to the Royal Society.

In the years since Thomas’ original work, his theorem has become ever-more significant. For an illustrated explanation of Bayes’ Theorem see The Less Wrong blog. As noted in Wikipedia, Bayes’ theorem is fundamental to probability theory, itself “a mathematical foundation for statistics, probability theory is essential to many human activities that involve quantitative analysis of large sets of data”. Now we are on home ground: creating value from big data.

So today we give praise to Bayes in general and the specific. In general, for working with others and creating the public space where today, our ideas thrive. Our assumption of access to this space is a defining characteristic of our modern age; echoes of our rage still echo from those days in late January when the Egyptian State attempted to isolate its citizenry from this public domain by blocking their access.

Specifically we praise Thomas Bayes for his theorem. Bill Bryson, author of At Home, an entertaining and sobering investigation of modern domesticity, captures the genius of the man: “The most remarkable feature of Bayes’s theorem is that it had no practical applications without computers to do the necessary calculations”. With computers we have applications that classify email as spam, find a missing US Navy submarine, forecast the weather, and predict stock and currency movements on financial markets. Many within IBM and Netezza believe these applications mark the beginning of a journey, that far greater value will be generated as we pair thoughtfully-designed computer systems with statistical analyses such as those based on Bayes’ theorem.

Thomas’s birth date appears lost to history, so we choose today, April 17, to praise the Reverend Thomas Bayes.

I recommend this video of Bill Bryson’s 2010 lecture marking the 350th anniversary of The Royal Society in which he expresses his admiration for Thomas.

Also please check out our video series that details Bayes effects on modern analytics:

  1. The man behind the Bayesian Theorem
  2. Thomas Bayes and Modern Intelligence Gathering
  3. Thomas Bayes keeps your inbox spam-free


Download Whitepaper: 

The Netezza Data Appliance Architecture

A Platform for High Performance Data Warehousing and Analytics