I heard a fascinating piece on the radio this morning about using statistics to help find missing aircrafts. Of course it was part of the flight 340 coverage, but the specific case they cited was the 2009 Air France plane that disappeared in the South Atlantic. It seems that using Bayes' Theorem, a statistical model that can compute a likely outcome when there are many competing variables (like where a missing plane might be), can be helpful.
Coincidentally, I also read an article today about a company that plans to use the data gleaned by tracking kids' responses to computer programs to develop a complete educational profile and action plan for every student. Education happens to be today the most data-mineable industry by far,” says their CEO in this video. "Every single thing in education is correlated to something else."
No doubt, they, too plan to use some iteration of Bayes' ideas to develop their automated response to students' needs. To many, that approach may sound ideal, but Arnold Barnett, a statistician at MIT, included this disclaimer in the radio piece this morning, "Bayes Theorem can't find the plane, period. It can, at best, change the odds."
And in fact, when they applied the theorem to the Air France flight five years ago, they "eliminated huge swaths of ocean floor because nobody heard a signal from the plane's black boxes. But it turned out, against the odds, both of the black boxes were damaged." It took two years to find that plane.
In the words of Colleen Keller, the mathematician working on that case, "Sometimes the probabilities will turn around and bite you."
Look out kids!
Coincidentally, I also read an article today about a company that plans to use the data gleaned by tracking kids' responses to computer programs to develop a complete educational profile and action plan for every student. Education happens to be today the most data-mineable industry by far,” says their CEO in this video. "Every single thing in education is correlated to something else."
No doubt, they, too plan to use some iteration of Bayes' ideas to develop their automated response to students' needs. To many, that approach may sound ideal, but Arnold Barnett, a statistician at MIT, included this disclaimer in the radio piece this morning, "Bayes Theorem can't find the plane, period. It can, at best, change the odds."
And in fact, when they applied the theorem to the Air France flight five years ago, they "eliminated huge swaths of ocean floor because nobody heard a signal from the plane's black boxes. But it turned out, against the odds, both of the black boxes were damaged." It took two years to find that plane.
In the words of Colleen Keller, the mathematician working on that case, "Sometimes the probabilities will turn around and bite you."
Look out kids!
No comments:
Post a Comment