Prediction of Recidivism in Thefts and Burglaries Using Machine Learning

Prediction of Recidivism in Thefts and Burglaries Using Machine Learning
复制标题

使用机器学习预测盗窃和入室盗窃中的累犯

DOI：
10.17485/ijst/2020/v13i06/149853
复制
发表时间：
2020
期刊：
影响因子：
--
通讯作者：
Fredy Humberto Troncoso Espinosa
Fredy Humberto Troncoso Espinosa
中科院分区：
--
文献类型：
--
作者：
Fredy Humberto Troncoso Espinosa

文献摘要

被引文献

相似文献

背景/目标：盗窃和入室盗窃是两种具有巨大社会影响的财产犯罪。他们的预防大大降低了受害率和民众的不安全感。本次调查的目的是获得一个指数，可以预测犯罪分子在此类犯罪中的重复犯罪，以支持预防行动的决策。方法：为了获得该指数，一组机器学习人员接受了培训，信息由智利比奥比奥地区检察官办公室的犯罪分析和调查焦点系统 (CAIFS) 提供。所提供的信息来自 2012 年至 2017 年间康塞普西翁市发生的盗窃和入室盗窃案。研究结果/应用：结果显示了此类犯罪中累犯的特征以及累犯指数，与目前使用的方法相比，该指数在累犯预测方面具有更大的自信。
Background/objectives: Theft and burglary are two crimes against property that have a great social impact. Their prevention drastically lowers victimization rates and the feeling of insecurity in the population. The objective of this investigation is to obtain an index that allows the prediction of repeat offenses by criminals in these types of crimes, in order to support decision-making with respect to preventative actions. Methodology: In order to obtain the index, a group of machines learning was trained, with information provided by the Criminal Analysis and Investigative Focus System (CAIFS) from the Regional Public Prosecutor’s Office in Biobío, Chile. The information provided was from thefts and burglaries committed between 2012 and 2017 in the city of Concepción. Findings/application: The results show a characterization of repeat offenders in these types of crime and a recurrence index that allows for a greater assertiveness in the prediction of recidivism than the method that is currently being used.