As reported in Omnia and Penn Today, PSC Associate Richard Berk discusses how existing crime patterns can be used to forecast violent events. His research, conducted with Susan B. Sorenson of the School of Social Policy & Practice, uses a predictive algorithm and a large dataset to anticipate intimate-partner violence. “Ideally the same strategy would work for mass violence,” says Berk.
Richard Berk commented in a Wall Street Journal article on the insignificance of a less than 1 percent decrease in violent crime from 2016 to 2017, as reported by the FBI.
Richard Berk of the School of Arts & Sciences is quoted in a Washington Post article about crime that occurs on a local level. He says, "Some cities that have more problems than others, and in those cities some neighborhoods have more problems than others, and to talk about national anything is just politics."
Richard Berk and his co-investigator Susan Sorenson have found that using machine learning at arraignment proceedings for domestic violence offences can help reduce repeat offending and this research which appeared in the The Journal of Empirical Legal Studies was discussed by the Penn News in February and Omnia in June.
In an article on fivethirtyeight.com, "Should Prison Sentences Be Based On Crimes That Haven’t Been Committed Yet?" Richard Berk talks about how risk assessment is used to identify offenders who are more likely to re-offend, and in what ways.
Richard Berk comments on predicting crime in the Atlantic's article, "Misfortune Teller."
Richard Berk discusses his development of software that can predict criminal behavior on ABC News.