Research on Cybercrime Based on BP Neural Network Algorithm

Abstract

Cybercrime is a variety of criminal activities carried out using digital devices or networks. Owing to its inherent complexity, countries have developed different policies and laws to safeguard cybersecurity. This paper shows the global distribution of different indicators of cybercrime, examines the relevance of different national policies, and explores the extent to which different demographic characteristics influence cybercrime. This article first defines four main indicators to describe cybercrime, and collects data related to cybercrime from VCDB and NETSCOUT to display the global distribution of these four indicators. A model for extracting distribution characteristics of cybercrime was constructed, and Pearson coefficient was used to analyze the relationship between cybercrime and many influencing factors. Finally, it was concluded that the distribution of cybercrime is concentrated in countries with larger populations, weaker network protection technologies and infrastructure, and weaker judicial systems. At the same time, a network security policy correlation analysis model was constructed using the five network security policy indicators provided by ITU. The GUI indicators were combined with the distribution and data preprocessing of cybercrime, and linear regression algorithm was used to analyze the correlation of GUI policy indicators. By using the BP neural network algorithm to reduce the impact of nonlinear relationships, it was found that the formulation of intergovernmental cooperation policies and government capacity building policies has a more significant effect on suppressing cybercrime. Finally, this article analyzes the impact of specific national policies on cybercrime.