CDAPS – Crime Data Analysis & Prediction System

Abstract - Crime prediction is becoming increasingly pivotal for effective community safety and resource allocation. Contemporary research leverages data mining, spatio-temporal analytics, and machine learning models to forecast crime trends, identify hotspots, and suggest intervention strategies. This paper synthesizes methodologies and findings from recent works to present a nuanced framework for predicting criminal activity, integrating supervised learning, clustering, and deep learning models. With a focus on both administrative and alternative data sources, particularly large-scale news articles, the study demonstrates methods for improving prediction accuracy, balancing datasets, and ensuring actionable insights for law enforcement agencies.

Key Words: machine learning, crime prediction, spatio- temporal analytics, data mining, classification, random forest.