Predicting recidivism using Machine Learning

What is it about?

This study aims to improve the predictive performance of existing risk assessment tools and the predictive validity of the original Ministry of Justice Case Assessment Tool (MJCA) concerning recidivism rates using machine learning (ML) and examine whether the tool’s predictive performance can be improved. With follow-up data on 5,942 individuals in Japanese Juvenile Assessment Centers, the study uses ML methods, such as the K-nearest neighbor algorithm, support vector machine, random forest, gradient boosting tree, and multilayer perceptron, to improve the MJCA’s prediction power. The results show that the predictive validity of the original MJCA significantly improves for three of the six ML methods; gradient boosting tree, random forest, and multilayer perceptron have the highest predictive validity.

Featured Image

Photo by Andy Kelly on Unsplash

Photo by Andy Kelly on Unsplash

Why is it important?

The importance of this study lies in the fact that it indicates the benefits and limitations of machine learning through comparing and testing the performance of several types of machine learning on predicting recidivism in a large sample of cases. We concluded that ML could improve the predictive validity of recidivism rates. The improvement is less pronounced than the enormous impact that recent artificial intelligence methods have had on information processing. However, it is significant because recidivism risk assessment is important in determining the treatment for individuals who offend. ML is beneficial for risk assessment and must be used with a focus on these issues.

Perspectives