Thursday, January 27, 2011

Predictions for Juárez

As long as we're on the subject of predictions, La Plaza has a piece about a Juárez academic team with a model that predicted 3,000 murders in 2010, and is predicting 5,000 killings in 2011:
An artificial-intelligence model generated by a university researcher in Ciudad Juarez, Mexico, projects that 5,000 people will be killed in the violent border city this year. The same model projected at the start of 2010 that 3,000 would be killed in the greater Juarez area, a figure that eventually reached 3,111 -- about a 94% accuracy rate.

It may seem far-fetched to make such long-term projections on a fluid criminal conflict such as the drug war in Juarez. Researcher Alberto Ochoa, in an interview with La Plaza on Monday, said his model is based on methods that mimic biology-based, or "bioinspired," patterns. Barring a "radical change" in Ciudad Juarez -- where the Juarez and Sinaloa cartels are battling over the drug-trafficking route across the U.S. border into El Paso, Texas -- his projection foresees a figure of roughly 5,000 dead.

"This technique is nothing new," Ochoa said from the Autonomous University of Ciudad Juarez, where he is a researcher at the Center for Social Investigations. "It's not the most accurate model but it is based heavily on reality."

"It's not Excel," the researcher added, referring to the commonly used software program. "The model has to be fed, values have to be adjusted. It's complicated."
I know nothing about the science, but the concept strikes me as being of limited value. He says barring a "radical change", but what about a bunch of semi-radical changes? In any event, what is life if not a series of radical and semi-radical changes?

No comments: