Analysis of the effectiveness of machine learning model in predicting the risk of inmates

doi:10.3969/j.issn.1000-1565.2017.04.015机器学习模型在预测服刑人员

Abstract

Abstract: By analyzing the current situation of risk assessment of inmate at home and abroad, we find that the adaptability and accuracy of the traditional risk assessment tool of inmate based on the scale is being in creasingly challenged.However,the machine learning model driven by the data and parameter can be self learning,so as to continuously improve the applicability and accuracy of the model.Firstly, the paper introduces the four common machine learning models of LR, CART, CHAID and MLPNN; then,using the 2004 survey of inmates in state and federal correctional facilities(SISFCF)as the data source, the four models were evaluated by the sensitivity, specificity, accuracy, AUC and other evaluating indicators;finally, the predictive ability of the four models are compared.

Key words: machine learning, prediction, recidivism, risk assessment

CLC Number:

TP393.08

MA Guofu,WANG Zixian,MA Shengli. Analysis of the effectiveness of machine learning model in predicting the risk of inmates[J]. Journal of Hebei University (Natural Science Edition), 2017, 37(4): 426-433.

References

[1] DAN W,BEECH A,BLACKER H F M J.Actuarial risk assessment and recidivism in a sample of UK intellectually disabled sexual offenders[J].Journal of Sexual Aggression,2009,15(1):97-106.DOI:10.1080/13552600802578577.
[2] WIJK L, EDELBRING S, SVENSSON A, et al.A pilot for a computer-based simulation system for risk estimation and treatment of mentally disordered offenders[J].Informatics for Health& Social Care,2009,34(2):106-115.DOI:/10.1080/17538150903014395.
[3] KELLY R,HELEN L M.The effect of training on the quality of HCR-20 violence risk Assessments in forensic secure services[J].Journal of Forensic Psychiatry&Psychology,2009, 20(3):473-480.DOI:10.1080/14789940802638366.
[4] GRANT D W.The development,validity and reliability of the minnesota screening tool assessing recidivism risk[J].Criminal Justice Policy Review,2014,25(5):579-613.DOI:10.11-77/0887403413478821.
[5] ZENGY J M,USTUNY B,RUDIN C.Interpretable classification models for recidivism prediction[EB/OL] (2015-03-30)[2016-8-16].http://arxiv.org/pdf/1503.07810v2.pdf.
[6] HAMILTON Z,NEUILLY M A,LEE S,et al.Isolating modeling effects in offender risk assessment[J].Journal of Experimental Criminology,2015, 11(2):299-318.DOI:10.1007/s11292-014-9221-8.
[7] HOCHSTETLER A, PETERS D J, DELISI M.Classifying risk development and predicting parolee recidivism with growth mixture models[J].American Journal of Criminal Justice,2016,41(3):602-620.DOI:10.1007/s12103-015-93-208.
[8] DUWE G, KIM K D.Out with the old and in with the new An empirical comparison of supervised learning algorithms to predict recidivism[J].Criminal Justice Policy Review,2015,41(9):1-31.DOI:10.1177/0887403415604899.
[9] 胡庆生.行刑方式的文明进步-上海市积极拓展社区矫治新空间[N].法制日报,2003-08-04(8).
[10] 黄兴瑞,孔一,曾贇.再犯预测研究-对浙江罪犯再犯可能性的实证分析[J].犯罪与改造研究,2004(8):8-13. HUANG X R,KONG Y,ZENG Y.Prediction of recidivism-empirical analysis of the possibility of recidivism in Zhejiang[J].Research on crime and transformation,2004(8):8-13.
[11] 邬庆祥.刑释人员人身危险性的测评研究[J].心理科学,2005,28(1):222-224.DOI:10.16719/j.cnki.1671- 6981.2005.01.063. WU Q X.A research on the appraisal of the personal dangerousness of persons released after completion of a sentence[J].Psychological Science,2005,28(1):222-224.DOI:10.16719/j.cnki.1671-6981.2005.01.063.
[12] 周志华.机器学习[M].北京:清华大学出版社,2016.
[13] 曾赟.服刑人员刑满释放前重新犯罪风险预测研究[J].法学评论,2011(6):131-137.DOI:10.13415/j.cnki.fxpl.2011.06.003. ZENG Y.Prediction of risk of redivism before the offenders released from prison[J].Law Review,2011(6):131-137. DOI:10.13415/j.cnki.fxpl.2011.06.003.
[14] 孔一,黄兴瑞.刑释人员再犯风险评估量表(RRAI)研究[J].中国刑事法杂志,2011(10):91-106. KONG Y,HUANG X R.Study of recidivism risk assessment list for released offenders[J].Journal of Chinese criminal law,2011(10):91-106.
[15] 丁传庆,鲁兰,任军,等.中国监狱罪犯分类理论与实务研究[J].刑事法评论,2015(1):636-652. DING C Q,LU L,REN J.Research on the theory and practice of the classification of prison criminals in china[J].Criminal law review,2015(1):636-652.
[16] WITTEN I H,FRANK E,HALL M A.数据挖掘:实用机器学习工具与技术(原书第3版)[M].北京:机械工业出版社,2014.a.m.
[17] 李航.统计学习方法[M].北京:清华大学出版社,2015.
[18] FELLOW S T, LEESE M.A green-fingered approach can improve the clinical utility of violence risk assessment tools[J].Criminal Behaviour & Mental Health,2003,13(3):153-158.DOI: 10.1002/cbm.538.
[19] 杨友星.CHAID 算法并行化及其在信用风险分析中的应用[D].长春:长春工业大学,2016. YANG Y X.CHAID Algorithm parallelization and application in credit risk analysis[D].Changchun:Changchun University of Technology,2016.
[20] GOTTFREDSON S D, MORIARTY L J.Statistical risk assessment: old problems and new applications[J]. Crime&Delinquency, 2006, 52(52):178-200.DOI: 10.1177/0011128705281748.
[21] HAN J W,KAMBER M, PEI J.数据挖掘概念与技术(第3版)[M].北京:机械工业出版社,2015.