AI, Recidivism Prediction, XGBoost, K-Means, Prisoner Rehabilitation
-
Home
-
Activities & Events
-
Master’s Thesis Defense on Predicting Recidivism Risk Using Artificial Intelligence in Marib
Master’s Thesis Defense on Predicting Recidivism Risk Using Artificial Intelligence in Marib
The Arab Academy for Management, Banking, and Financial Sciences held the public defense of a Master’s thesis submitted by researcher Ali Ali Naji Abdullah Farhan as part of the Master’s Program in the Department of Computer Information Systems. The thesis was entitled:
“Prisoner Rehabilitation Based on Recidivism Risk Prediction Using K-Means and XGBoost Algorithms: A Case Study of Marib Governorate.”
The study aimed to develop an intelligent framework to support prisoner classification and predict the risk of recidivism by employing artificial intelligence and machine learning techniques. The framework is intended to support the design and allocation of rehabilitation programs according to risk levels and individual needs.
The study utilized a dataset containing 3,111 records from the Rehabilitation and Correction Authority in Marib Governorate covering the period from 2021 to 2025. It integrated the XGBoost algorithm for prediction, the SMOTE technique to address class imbalance, and SHAP analysis to interpret the model’s results, alongside the K-Means algorithm to classify prisoners into seven behavioral groups based on the Risk-Need-Responsivity (RNR) model.
The results demonstrated high performance indicators for the proposed model, with an overall accuracy of 99.19% and a ROC-AUC of approximately 0.9998. SHAP analysis also indicated that criminal history and annual crime rate were among the most influential factors in the predictions.
The Thesis Examination Committee consisted of Associate Professor Dr. Mukhtar Hussein Abdulwahab Al-Sarouri, Chair and External Examiner; Dr. Abdulaziz Ahmed Abdulaziz Thawaba, Supervisor and Member; and Dr. Yahya Mohammed Ali Mohammed, Internal Examiner and Member.
At the conclusion of the session, the committee announced its decision to accept the thesis and recommend awarding researcher Ali Ali Naji Abdullah Farhan a Master’s degree in Computer Information Systems, subject to incorporating the proposed amendments. The committee wished him continued success in his academic and professional career.
Share this post:
Related Posts