CybercrimeIR - A Technological Perspective to Fight Cybercrime

摘要

The problem of cybercrime is so serious and costs our society increasingly and significantly. Integrating the cybercrime materials of law enforcement is urgent need for fighting cybercrime internationally. This study proposes a feasible architecture of CybercrimeIR to collect and classify the useful cybercrime materials in investigator's perspective. In the experiments, this study firstly adopts text representation approaches and machine learning techniques (e.g. support vector machine, Naive Bayesian, and C4.5) to classify useful cybercrime materials in investigators' perspective. The performance measure in accuracy can at least achieve to 90% while conducting feature selection with information gain. We believe the proposed architecture of CybercrimeIR is very useful for integrating cybercrime materials of law enforcement globally.