Finding the Limits of Bayes’ Rule by Searching for a Missing Plane
Bayes’ rule, the 250-year-old statistical theorem for calculating conditional probabilities, has been employed in the search team for Malaysia Airlines Flight 370. The article discusses how Colleen Keller, a senior analyst at Metron, has developed a “Bayesian” model to look for planes. She starts by compiling data of the flight’s last known position, details of weather and currents, and statistics from previous crashes. Then, she calculates the odds of different scenarios and combines them in a large probability framework. Finally, she uses the model to calculate the odds of the plane being at any one point in a nearly 50-mile radius of the last point of contact.
This method of search was first employed in the search for Air France Flight 447, which disappeared in 2009. However, it was ineffective because both “black boxes” within the planes were damaged, which gave no reference point for Keller’s algorithm to work from. Thus, although the method proves useful for researchers working in a vast area, it definitely has its limits. “If the search itself is imperfect, then the revisions under Bayes’ Theorem will still leave a great deal of uncertainty,” says Keller. All in all, Keller’s employment of Bayes’ Theorem in the search for missing planes highlights how limited algorithms can offer an excellent basis for procedure, but often require the power of luck to achieve the desired solution.
https://www.npr.org/sections/thetwo-way/2014/03/25/294390476/can-a-250-year-old-mathematical-theorem-find-a-missing-plane