Abstract
The regression model and random forest model have been used to research recidivism prediction. The purpose of the research is firstly, to determine the crucial factors influencing recidivism rates among gender, ethnicity, prior crime record, misdemeanor status, and age. The second purpose is to discuss potential bias among the raw data. The third one is determining the pros and cons of machine decisions. With the work of data processing, analysis, and modeling, we had come to our result that factor terms that have a significant impact on the recidivism rate include: gender, priors, priors interact with age range, gender interact with age range. Through a variety of modeling methods, we established logistic regression models and random forest models to predict the crime rate through variables such as gender, age, and so on. Both models show decent accuracy, and two different models can be adapted to different situations. Some limitations have been admitted after comparison to other papers, such as the usage of only static variables, and at the same time, potential future improvements regarding the work are proposed in accordance with the limitations found.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call