Do Machine Learning Methods Outperform Traditional Statistical Models in Crime Prediction? A Comparison Between Logistic Regression and Neural Networks
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Although machine learning (ML) methods have recently gained popularity in both academia and industry as alternative risk assessment tools for efficient decision-making, inconsistent patterns are observed in the existing literature regarding their competitiveness and utility in predicting various outcomes. Drawing on a sample of the general youth population in the U.S., we compared the predictive accuracy of logistic regression (LR) and neural networks (NNs), which are the most widely applied approaches in conventional statistics and contemporary ML methods, respectively, by adopting many theoretically relevant predictors of the future arrest outcome. Even after fully implementing rigorous ML protocols for model tuning and up-sampling and down-sampling procedures recommended in recent literature to optimize learning algorithms, NNs did not yield substantially improved performance over LR if we still rely on a conventional dataset with relatively small sample sizes and a limited number of predictors. Nonetheless, we encourage more rigorous, comprehensive, and diverse evaluation research for a complete understanding of the ML potential in predictive capacity and the contingencies in which modern ML methods can perform better than conventional parametric statistical models.
Abstract
Although machine learning (ML) methods have recently gained popularity in both academia and industry as alternative risk assessment tools for efficient decision-making, inconsistent patterns are observed in the existing literature regarding their competitiveness and utility in predicting various outcomes. Drawing on a sample of the general youth population in the U.S., we compared the predictive accuracy of logistic regression (LR) and neural networks (NNs), which are the most widely applied approaches in conventional statistics and contemporary ML methods, respectively, by adopting many theoretically relevant predictors of the future arrest outcome. Even after fully implementing rigorous ML protocols for model tuning and up-sampling and down-sampling procedures recommended in recent literature to optimize learning algorithms, NNs did not yield substantially improved performance over LR if we still rely on a conventional dataset with relatively small sample sizes and a limited number of predictors. Nonetheless, we encourage more rigorous, comprehensive, and diverse evaluation research for a complete understanding of the ML potential in predictive capacity and the contingencies in which modern ML methods can perform better than conventional parametric statistical models.
참고문헌 (33)
- Berk, R. A., "Criminal justice forecasts of risk: A machine learning approach" (2012), Springer
- Richard A. Berk, "Statistical Procedures for Forecasting Criminal Behavior", Criminology & Public Policy, 12(3), pp.513~544 (2013), Wiley[DOI]
- Bishop, C. M., "Pattern Recognition and Machine Learning" (2009), Springer
- Breiman, L., "Decision tree forest", Machine Learning, 45(1), pp.5~32 (2001)
- Brodzinski, J. D., "Using artificial intelligence to model juvenile recidivism patterns", Computers in Human Services, 10, pp.1~18 (1994)
- "Using offender risk and needs assessment information at sentencing: Guidance from a national working group" (2011)[URL]
- Caulkins, J., "Predicting the criminal recidivism : A comparison of neural network models with statistical models", Journal of Criminal Justice, 24, pp.227~240 (1996)
- Farrington, D. P., "Crime and justice: an annual review of research (Vol. 9)" (1987), University of Chicago Press
- Farrington, D. P., "Prediction in criminology" (1985), State University of New York Press
- Gottfredson, S. D., "Screening for risk: A comparison of methods" (1979), U.S. Government Priming Office
- Stephen D. Gottfredson, "Screening for risk: A comparison of methods", Criminal Justice and Behavior, 7(3), pp.315~330 (1980), SAGE Publications[DOI]
- Gottfredson, S. D., "Criminal careers and “career criminals” Vol. 2", pp.212~290 (1986), National Academy of Sciences Press
- Gottfredson, S. D., "Statistical risk assessment : Old problems and new applications", Crime & Delinquency, 52, pp.178~200 (2006)
- Hastie, R., "Rational choice in an uncertain world : The psychology of judgment and decision making. Sage Publications", Sage Publications, Inc
- Jamain, A., "Mining supervised classification performance studies : A meta-analytic investigation", Journal of Classification, 25, pp.87~112 (2008)
- Kartalopoulos, S. V., "Understanding Neural Networks and Fuzzy Logic: Basic Concepts and Applications" (1995), IEEE Press
- Kuhn, M., "Applied Predictive Modeling" (2013), Springer-Verlag
- Liu, Y. Y., "A comparison of logistic regression, classification and regression trees, and neutral networks model in predicting violent re-offending", Journal of Quantitative Criminology, 27, pp.547~573 (2011)
- Christopher T. Lowenkamp, "Risk", Need Assessment, Offender Classification, and the Role of Childhood Abuse, 2001, pp.28(5), 543~563[DOI]
- Mossman, D., "Assessing predictions of violence: Being accurate about accuracy", J. Consultng Clin. Psychol, 6, pp.783~792 (1994)
- Fawn T. Ngo, "Assessing the Predictive Utility of Logistic Regression, Classification and Regression Tree, Chi-Squared Automatic Interaction Detection, and Neural Network Models in Predicting Inmate Misconduct", American Journal of Criminal Justice, 40(1), pp.47~74 (2015), Springer Science and Business Media LLC[DOI]
- Ngo, F. T., "Traditional Regression Methods versus the Utility of Machine Learning Techniques in Forecasting Inmate Misconduct in the United States : An Exploration of the Prospects of the Techniques", International Journal of Criminal Justice Sciences, 13(2), pp.420~437 (2018)
- Palocsay, S. W., "Predicting criminal recidivism using neural networks", Socio-Economic Planning Sciences, 34, pp.271~284 (2000)
- "Risk/needs assessment 101: science reveals new tools to manage offenders" (2011)[URL]
- "R: A language and environment for statistical computing" (2018)[URL]
- Rumelhart, D. E., "Parallel distributed processing: Explorations in the micro-structure of cognition (Vol. 1–2)" (1986), The MIT Press
- Edie S. Sears, "Artificial Neural Networks for Adolescent Marijuana Use and Clinical Features of Marijuana Dependence", Substance Use & Misuse, 39(1), pp.107~134 (2004), Informa UK Limited[DOI]
- Jennifer L. Skeem, "Current Directions in Violence Risk Assessment", Current Directions in Psychological Science, 20(1), pp.38~42 (2011), SAGE Publications[DOI]
- Tollenaar, N., "Which method predicts recidivism best? A comparison of statistical, machine learning and data mining predictive methods", Journal of the Royal Statistical Society, Series A, 176(part 2), pp.565~584 (2013)
- Van Voorhis, P., "Risk classification in the 1990s" (1990), National Institute of Corrections
- Vapnik, V., "The Nature of Statistical Learning Theory" (2010), Springer
- Venables, W. N., "Modern Applied Statistics with S" (2002), Springer
- Jiaming Zeng, "Interpretable classification models for recidivism prediction", Journal of the Royal Statistical Society: Series A (Statistics in Society), 180(3), pp.689~722 (2017), Wiley[DOI]