An algorithm has appeared that predicts crimes in a week

The new algorithm studies patterns in the time of crimes and their location based on

open data on violent and property crimes. The model predicts crimes a week before and with an accuracy of about 90%.

Within a separate model, researchthe team also studied the response of the police to the crimes by analyzing the number of arrests after the incidents and comparing these rates among districts with different socioeconomic status. They saw that crime in wealthier areas led to more arrests, while fewer arrests were made in disadvantaged areas. This speaks to police bias.

The tool was tested in Chicago for two categories of crimes:

  • violent crimes (murders, assaults and beatings),
  • property crimes (burglaries, thefts and car thefts).

The team exploited these offences, as they are more likely to be reported to the police. Such crimes are also less susceptible to law enforcement bias.

The new model separates crimes from one another. It looks at the temporal and spatial coordinates of individual events to find patterns for predicting the future.

The model divides the city into sectors and predicts the level of crime within these boundaries instead of using traditional boundaries.

The team notes with caution that the accuracytool does not mean that it should be used as a guide for law enforcement. Instead, it is advised to add it to the toolkit of city policy and police strategies to combat crime.

Read more:

Astronomers have calculated when the planets will change their orbits and chaos will come

The space probe flew 200 km from Mercury. Look what he saw

The unique spaceplane will deliver cargo to the ISS. He is not like the others