Research on Cybercrime based on K-means Clustering Algorithm and Linear Regression Modelling

Abstract

This paper focuses on research related to cybercrime governance, aiming to analyse the distribution of cybercrime globally, the factors influencing it and the role of different policies in curbing it. Through data collection and visual analysis, significant differences were found between countries on five dimensions of cybercrime governance, namely legal, technological, organisational, capacity development and cooperation measures. The K-means clustering algorithm was used to classify countries into four risk clusters, which showed that some African countries were in the high-risk group, while developed countries were mostly in the low-risk group. After Pearson correlation analysis, a strong positive correlation was observed among the governance measures. Linear regression modelling using the least squares method, with cybercrime incidence as the dependent variable and policy scores and pattern scores as the independent variables, yielded that both pattern scores and policy scores were negatively correlated with cybercrime incidence, with pattern scores having a stronger impact. The study shows that most developing countries are less capable of curbing cybercrime, while developed countries have more comprehensive measures, and that cybercrime rates are negatively correlated with response scores, findings that provide an important basis for the formulation of cybersecurity policies.