Monday, April 07, 2008

Prediction is Hard: Pentagon on Social Networks

According to DangerRoom some people at Pentagon now try to predict (at least parts of the) future through some advanced social science:
The University of Pennsylvania School of Engineering and Applied Science just announced that it's been awarded a $7.5 million grant to work in this fast emerging field of network science, which melds everything from mathematics to sociology. Network science is increasingly the "hot" area for Pentagon research. Why? Because the Pentagon hopes that if it can understand complex networks, then it can understand terrorist networks, and even predict who will join such a network.
While there's a lot of good things to be said about attempts at modeling market behaviors through quantitative approaches, it is certainly wise to bring a pinch of reservation to the table as well. In one interesting attempt to forecast country instability, "Anticipating the Good, the Bad & the Ugly" (Journal of Conflict Resolution), Sean P. O'Brien admits that "the model can anticipate the oiliness of the rags but not the spark that will set them ablaze." Prediction, in other words will not be wrought from quantitative approaches. Trend forecasting and identification of likelihoods is what that game is about. Engineering, in this sense, as an organizational approach will only bring you so far. Leadership, intuition, horizontal thinking and perhaps common sense is still necessary.

No comments: