應用類神經網路建構妨害性自主罪再犯預測模型之初步嚐試

The Preliminary Attempt to Use Artificial Neural Network Model for the Sexual Offender Recidivism Prediction

劉昌誠(Chang-Cheng Liu) ； 徐建業(Chien-Yeh Hsu) ； 陳炯旭(Chiung-Hsu Chen) ； 蕭百勝(Bai-Sheng Hsiao)

《亞洲家庭暴力與性侵害期刊》 6卷1期 (2010/07) Pp. 43-64

https://doi.org/10.29804/AJDVSO.201007.0003

摘要

Participants of this study were 349 sexual offenders released from a prison in northern Taiwan in 1995, and we follow all cases from the time of release to December 31, 2003. 18 risk predictors with statistic significance are selected to construct a artificial neural network (ANN) model for the sexual offender recidivism prediction. Then we examined the predict ability of the ANN model by receiver operating characteristic (ROC) analysis, and compare with other common used screening tools for prediction of sex offender recidivism, i.e. RRASOR, static-99, and MnSOST-R. The area under the ROC curve for ANN model is 0.772 (95% CI: 0.683~0.862, p＜0.001), which reach the statistic significance. Comparing with other screening tools, ANN model got better predict ability. It means that ANN is a considerable method for further research.

關鍵字

妨害性自主 ； 再犯預測 ； 類神經網路

並列摘要

Participants of this study were 349 sexual offenders released from a prison in northern Taiwan in 1995, and we follow all cases from the time of release to December 31, 2003. 18 risk predictors with statistic significance are selected to construct a artificial neural network (ANN) model for the sexual offender recidivism prediction. Then we examined the predict ability of the ANN model by receiver operating characteristic (ROC) analysis, and compare with other common used screening tools for prediction of sex offender recidivism, i.e. RRASOR, static-99, and MnSOST-R. The area under the ROC curve for ANN model is 0.772 (95% CI: 0.683~0.862, p＜0.001), which reach the statistic significance. Comparing with other screening tools, ANN model got better predict ability. It means that ANN is a considerable method for further research.

並列關鍵字

Sexual offense ； prediction of recidivism ； artificial neural network