Data Mining and Machine Learning Approaches for Analyzing Drug-Related Overdose Patterns and Risk Factors

This study applies cutting-edge data mining and machine learning techniques to examine critical determinants of drug overdose death. Through the analysis of overdose-related data sets, the study aims to identify high-risk age groups, demographic clusters, and spatial mobility patterns associated with fatal drug consumption, namely fentanyl, fentanyl analogues, and xylazine. Clustering algorithms such as KMeans and HDBSCAN are used to detect hidden demographic and geographic patterns, while classification models such as Random Forest, Support Vector Machine, and K-Nearest Neighbors are employed to estimate substance use and identify major risk factors. Anomaly detection techniques are also employed to investigate outlier geographic displacement in overdose cases, which can offer insight into drug trafficking and drug tourism tendencies. The findings will guide evidence-based intervention strategies, strengthen the public health response, and guide policy to reduce the burgeoning drug overdose epidemic.