The Los Angeles Police Department and the UCLA Department of Mathematics have teamed up to use quantitative analysis and computer programming to help solve ongoing criminal investigations.
The project is in its final testing phases. It is the ultimate product of several years of gathering data and a research project begun by an undergraduate student, said Martin Short, one of the algorithm’s creators and a mathematics professor.
Police will be able to use the program to determine the probability of a gang’s involvement in a crime, he said.
Investigators input recent criminal activity by classifying events by individuals, their gang affiliations and the crime committed. After this, certain patterns emerge based on peoples’ predictable actions, such as retaliation for a shooting, said Andrea Bertozzi, another creator of the algorithm and an applied mathematics professor.
The program draws connections between these events ““ such as gunshots fired during a crime and the suspect’s prior criminal record ““ that a human investigator might miss, she said.
“Police rely a lot on maps on a wall and drawing connections on looking at the location of crimes,” Bertozzi said. “The algorithm is a more sophisticated method for investigation.”The project is an attempt to reconcile a growing amount of recorded data police collect with common investigatory practices, she said.
Technology such as the UCLA algorithm is crucial to helping catch criminals and prevent future crimes, said LAPD Captain Sean Malinowski.
“We’re trying to get out in front of crime instead of chasing after it,” he said. “UCLA and other schools have the skill set to help us improve our tactics.”
He added that it is important that police technology stays ahead of criminals’ activity because using mathematical models has been part of the reason Los Angeles crime has dropped about 15 percent in the past four years.
But the program does not reduce the need for human skill and intuition.
“(The algorithm) can only determine which gang a criminal might belong to based on data, not the individual who actually did it,” Bertozzi said.
“This isn’t a replacement for good old-fashioned police work.”
So far, the program has primarily been tested in theoretical situations and was capable of determining the top three gangs most likely involved with a crime with 80 percent success, Short said.
Police in Santa Cruz also have tested the system for the first time with crimes they had already solved, and the system verified their findings, Malinowski said.
It is still unclear, however, whether its success in Santa Cruz will transfer to Los Angeles.
Santa Cruz has only about 50 police officers and 12 square miles to patrol, while Malinowski’s unit has 250 officers and 50 square miles to cover, he said.