Crime Link Prediction Across Geographical Location Through Multifaceted Analysis: A Classifier Chain Temporal Feature-Data Frame Joins

Crime link prediction across geographical locations is vital for law enforcement to uncover hidden connections between crime spanning different areas. Traditional  methods often fails to capture the complexity and temporal dynamics of crime data, limiting g their predictive power. This research introduces a novel approach to enhance crime link prediction by leveraging multifaceted analysis  that integrates multiple inputs and outputs. A classifier chain transformation is used for sequential multi-label classification, capturing interdependence between crime types across locations. The method  facilitates a comphrensive understanding of crime patterns over time. Experiment conducted on South Africa Police  Services  (SAPS) crime dataset demonstrate the proposed model's superior performance compared to state-of-the-art methods, achieving precision, recall, F1-score, and accuracy of 0.98, 0.99,0.99, and 98.99%, respectively. This research aims to contribute to crime link prediction model's, offering a more nuanced and robust framework for forensic experts and law enforcement.