Written by: STATISTICA 9/30/2009 4:19 PM
I enjoy watching The Daily Show with Jon Stewart. This 30-minutes comedy show focuses on current events. Jon is very skillful at interviewing fascinating people on complex subjects. Jon touches on core truths while using humor.
On Monday September 28, Jon Stewart interviewed "a master of game theory", Bruce Bueno De Mesquita. Bruce recently wrote a book on game theory titled The Predictioneer's Game: Using the Logic of Brazen Self-Interest to See and Shape the Future.
I had to look up "game theory" to make that I understood it. This is my favorite definition from 6 Degrees + 1 Game Theory = Social Network Analysis:
Game theory attempts to mathematically capture behavior in strategic situations, in which an individual's success in making choices depends on the choices of others. While initially developed to analyze competitions in which one individual does better at another's expense (zero sum games), it has been expanded to treat a wide class of interactions. Today, Game Theory is a sort of an umbrella or 'unified field' theory for the rational side of social science.
The Daily Show with Jon Stewart had a nice discussion on why people are uncomfortable with predictive modeling/data mining. If you can't watch the Hulu video above, here are the interesting quotes:
Jon Stewart (comedian): The human animal.. Is it disturbing that we can be plugged into a computer model and thus have 90% accuracy on our behavior... Bruce Bueno De Mesquita (book author): It does bother a lot of people. A lot people don't like the idea that we can predict. A lot of people think that you can't reduce us to a mathmatical equation. I would point out that if you throw a frisbee up in the air to a dog, it figures out where to jump and catch the frisbee. And if you ask a physicist to write down the equations, he could. I don't think the dog actually recongizes the equations, but acts just as if it made those calculations. So we [human beings] can be represented that way.
Jon Stewart (comedian): The human animal.. Is it disturbing that we can be plugged into a computer model and thus have 90% accuracy on our behavior...
Bruce Bueno De Mesquita (book author): It does bother a lot of people. A lot people don't like the idea that we can predict. A lot of people think that you can't reduce us to a mathmatical equation.
I would point out that if you throw a frisbee up in the air to a dog, it figures out where to jump and catch the frisbee. And if you ask a physicist to write down the equations, he could. I don't think the dog actually recongizes the equations, but acts just as if it made those calculations. So we [human beings] can be represented that way.
On a personal level, I feel uncomfortable being reduced to a number in a statistical analysis... in a predictive model. The dog/frisbee metaphor is a little too simple. The dog isn't worried about how the predictions will be used. The dog isn't concerned about receiving junk email or phone calls.
But Mesquita's point is also very valid. It is related to my Data Mining, The Nature of Reality. "Reality" can be described in any number of ways. Different situations require using different "ways" of seeing reality. Governments and companies can use predictive modeling to decrease fraud and target marketing (reducing mail to us.. the non-targets).
This interview made me start to wonder about game theory and data mining. Are people using these theories together? So I turned my friends google and bing.
The answer is "yes". I found several recent academic papers on this idea. For example, Combining data mining and Game Theory in manufacturing strategy analysis.
Is anyone working with game theory and data mining?
Image Credit
/aw
0 comment(s) so far...