Researchers at the University of Chicago have developed a new algorithm capable of determining the location and time of a crime in a major city, approximately 7 days before it happens. Yes Yes !
Algorithm 90% reliable (!)
With a reliability of 90% (!), the algorithm divides the city into zones of 92 m2 for greater accuracy. The researchers fed their artificial intelligence program with thousands of historical data on violent crimes and break-ins using Chicago as a test city, in order to detect recurring patterns over time for each of these areas.
Chicago, Los Angeles and Philadelphia, test cities
The tests also worked in other cities such as Los Angeles and Philadelphia. Namely, other algorithms of this type already existed, but this one stands out for its lack of bias. Indeed, previous models considered that crimes emerge from specific places and spread. An approach that ignores the social environment of cities, as well as the nuanced link between crime and the effects of police actions.
Better than Minority Report
Where previous programs tried to identify potential criminals (Minority Report style) and victims with a relatively low success rate, this new tool uses hundreds of thousands of sociological models to determine the risk of crime in a particular location , at a particular time. Source: Nature Human Behavior
