The algorithm predicts whether a growing child can become a criminal

Predicting whether a child is capable of committing an offense before age 18 will help us take timely preventive measures.

Perhaps many of us do not feel alien to the science fiction film called Minority Report . The film is about the inspector character John Anderton who is full of responsibility and is very confident in the Prediction Department. But it is ironic that one day the Department announced that John was the next murderer, the victim was a man he did not know. No other choice, he had to flee.

It is on film, where people can fully predict or know someone who will become a criminal in the future. What about real life? Now, crime researcher Richard Berk of the University of Pennsylvania in the United States is trying to make things seem impossible.

Scientists are trying to predict whether growing up children become criminals.(Photo Shutterstock / Tomsickova Tatyana).

Currently, Richard Berk has asked the Norwegian government to collect a large amount of data about this country's citizens, then combine all that data with a unique identification file. Berk will focus on studying files about children and their parents, especially they will focus on the living environment, the time of birth of the child so that it can be predicted before the age of 18, These children are capable of doing something illegal or not.

Richard Berk's inspection process is based on machine learning techniques (data analysis methods will be automated). This technique will include data that scientists design into algorithms, helping computers know how to identify special patterns in large data files, even including data sets that they have never seen. before.

Once the computer can identify the sample, it will combine with the previous findings to predict the result, to prevent criminal activity. Algorithms used in machine learning techniques can accurately predict a child's later behavior if the amount of information collected is high. However, this prediction is not an easy operation. If this experiment of Berk is successful, it is likely that many people will be arrested and sentenced before they commit bad behavior.

Previously, Richard Berk developed computer software to help predict whether the individuals under probation or parole could recidivism. Predicting this problem must also be based on the number of previous arrests by a person, the age of the first arrest, and the type of offense.

Berk said: "People often think that if a person ever committed a murder, in the future it may be repeated. But more importantly, it's the person's age. "The page at the age of 14 is a good predictor. However, if the person is guilty at age 30, there is not much possibility of forecasting."

Previously, Richard Berk developed computer software to help predict whether the individuals under probation or parole could recidivism.(Illustration).

Several years ago, researchers looked at data on more than 60,000 different criminals, including murder crime. They use an algorithm to find a set of people who are able to commit murder when pardoned or challenged. The Upenn team can identify 8 out of 100 criminals who can kill people in the future.

These initiatives will help police and people to identify who is at risk of becoming a dangerous element to the community. At the same time, if it is recognized that the subject in question is likely to recidivism if it is forgiven, the authorities will have a timely settlement.

This research has been published in The Conversation.