Testing the utility of the neural network model to predict history of arrest among intimate partner violent men

Abstract

Risk assessments are typically based on retrospective reports of factors known to becorrelated with violence recidivism in simple linear models. Generally, these linear models use onlythe perpetrators' reports. Using a community sample of couples recruited for recent male-to-femaleintimate partner violence (IPV; N = 97 couples), the current study compared non-linear neuralnetwork models to traditional linear models in predicting a history of arrest in men who perpetrateIPV. The neural network models were found to be superior to the linear models in their predictivepower. Models were slightly improved by adding victims' report. These findings suggest that theprediction of violence arrest be enhanced through the use of neural network models and by includingcollateral reports.