Synthetic Dataset for Detecting the Relationship Between Psychological Illnesses and Criminal Tendency Patterns Using Machine Learning

This record contains synthetic datasets generated for the study titled “Detecting the Relationship Between Psychological Illnesses and Criminal Tendencies Using Machine Learning.” The datasets were created for research and machine learning modeling purposes and do not contain any real personal, clinical, forensic, or identifiable individual data. The first dataset contains approximately 1,000 records, while the second dataset contains 10,000 records generated through a Gaussian Mixture Model-based data augmentation process. The datasets include variables such as psychological disorder, crime type, offender status, age, gender, crime location, and time. These variables were structured to support the analysis of psychological illness and criminal tendency patterns using machine learning methods. The datasets were used to compare Naive Bayes, K-Nearest Neighbor, Decision Tree, Support Vector Machine, and Logistic Regression algorithms, and to support explainability analyses such as SHAP, LIME, feature importance, ablation study, sensitivity analysis, and model generalization evaluation. These datasets are intended only for academic research, experimental modeling, and decision-support framework development. They should not be used for clinical diagnosis, legal judgment, automated decision-making, or real-world criminal profiling.