Home News Predictive policing substantially reduces crime in Los Angeles during months-long test

Predictive policing substantially reduces crime in Los Angeles during months-long test

Screen Shot 2015-10-13 at 8.52.32 AM
Credit: LAPD

A joint study by UCLA scholars and law enforcement officials suggests that mathematics can help reduce crime.

During the study, the group devised a mathematical model that was used to guide the Los Angeles Police Department where to deploy officers. Using the model, the LAPD was able to substantially lower crime during a 21 month period.

Jeffrey Brantingham, a UCLA professor and senior author of the study, said, “Not only did the model predict twice as much crime as trained crime analysts predicted, but it also prevented twice as much crime.”

A paper about the study was published online by the Journal of the American Statistical Association.

The LAPD had so much success using the model and decided to use it in 14 of the department’s 21 divisions, which is up from just three divisions in 2013.

According to UCLA Newsroom, the mathematical model was developed using six years of mathematical research and a decade of police crime data. It predicts the times and places where serious crimes will occur based on the historical crime data of the area.

Brantingham said one of the keys to its success is the algorithm behind it, which learns over time.

“In much the same way that your video streaming service knows what movie you’re going to watch tomorrow, even if your tastes have changed, our algorithm is constantly evolving and adapting to new crime data,” he said.

Starting in 2011, the researchers analyzed crime data and trends in the LAPD’s Southwest division and in two Kent divisions. This was done to determine if the model could predict when and where major crimes would occur in real time.

In Los Angeles, the analysis focused on burglaries and car thefts. In Kent, England, where they performed a similar experiment, they studied the same patterns in addition to violent crimes.

The researchers pit the computer model against crime analysts to see which could predict where crimes would happen more accurately.

The model correctly predicted the locations of crimes on 4.7 percent of its forecasts in Los Angeles, while the human analysts were correct on 2.1 percent of their forecasts.

In the two Kent divisions, the model accurately predicted the locations of crimes on 9.8 and 6.8 percent of its forecasts. Human analysts correctly predicted the crime locations on 6.8 and 4 percent of their forecasts.

Based on the results they achieved using the mathematical model, researchers estimate that the LAPD can save up to $9 million per year.

Brantingham emphasized that the algorithm is intended to help police officers do their jobs better, and can’t be used to replace actual police work. He also believes that the model can be used effectively by police in cities worldwide.

If you have any problems viewing this article, please report it here.


Please enter your comment!
Please enter your name here