Using machine learning, research develops models and predicts high -risk gandelers.

Credit: Pixabay/CC0 Public domain
Federal -approved firearm dealers are important targets for restrictions and execution efforts aimed at reducing the supply of firearms to bypassing illegal markets, but measure the degree of dealers involved in illegal shifting firearms. It is difficult to do.
In a new study using machine learning, researchers have examined firearm transactions and criminal gun collection records from 2010 to 2021, and the maximum number and maximum number of guns collected by crime within one year after sales. I have identified a dealer that sells a percentage. Possibility of illegal acts by dealers or personal traders. The predictive model can support the target execution, which helps to confuse the flow of firearms to criminals.
This study is published in criminal science and public policies by researchers at the Davis School at the University of California (UC).
“Most gun criminals have not obtained firearms directly from the approved dealers, and most dealers are complaining of the law, but even a small number of negligently or damaged dealers are used in crime. You can contribute greatly to the supply of.
The firearm dealer can make a detour of gun guns into the crime market through practices, such as failing to sell to straw buyers and performed necessary background checks. Previous studies show that execution measures and litigation for dealers that measure laws can prevent these actions and reduce firearm flow to crime markets.
In this study, researchers have developed two prediction models using machine learning technology. The first dealer year is classified as the top 5 % of the annual crime gun sales (the number of gun sales collected by crime within one year after sales). The second is to identify the top 5 % dealer year based on the ratio of sales collected within one year. Both models had a strong identification performance, and the first model was particularly effective in identifying the highest risk dealer.
The model is generally better than a simpler regression and rule -based approach, emphasizing the value of the data adaptation model for prediction. The main prediction factors include the sale of criminal guns the previous year, the average age of the buyer, the proportion of “cheap” handguns, and the local gunfire and assault rate.
Many of the most predicted dealers have sold a large amount of guns in a very short “crime time”, but also sold crime guns for over the years. This suggests that a group of relatively small dealers can target execution and provide an oversized effect. With the citation or the cancellation of the license, and a consistent target of a high -risk dealer can help you enhance deterrent, promote compliance, and reduce the supply of guns to criminals.
“Our survey results shows that the comprehensive California firearm trading and the recovery data of criminal guns can be potentially useful for identifying high -risk retailers. Lacour says. “This type of identification can improve the efficiency and effectiveness of tests and execution efforts to prevent negligence or corrupt dealers.”
In this study restriction, the author may not be a dealer who sells many short -time guns, but the unimposable dealer may not be reflected in short -time criminal statistics. I point out that there is.
They also point out that California is also a state with strict gunshots and a wide range of dealer regulations, so the number of dealers that measure negligence or law in research may be lower than the states with less regulations. 。 However, the results of research are unique to California, but the consistency of risk factors over various jurisdictions and the context of regulations suggests that models can provide information to other state approaches. The authors say.
Details: Hannah S. Laquur et Al, identification high -risk firearm dealers: California, criminal science and public policy (2025). Doi: 10.1111/1745-9133.12692
Provided by the American Criminal Society
Quotation: Using machine learning, research has developed a model, and high-risk acquired from https://phys.org/2025-01-01-machine-gun-dealers.html on January 28, 2025 We predict the gun dealer (January 28, 2025) (January 28, 2025)
This document is subject to copyright. There is no part that is reproduced without writing permission, apart from fair transactions for private research and research purpose. Content is provided only by information.