Detection of Dark Web Threats Using Machine Learning and Image Processing

TL;DR

This paper explores detecting dark web threats, especially human trafficking, using image processing and machine learning techniques like SVM and logistic regression, highlighting the effectiveness of SVM in threat detection.

Contribution

It introduces a methodology combining image processing with machine learning models to identify dark web threats, demonstrating the superiority of SVM over logistic regression.

Findings

SVM outperforms logistic regression in accuracy

Significant insights gained from exploratory data analysis

Effective threat detection using image processing techniques

Abstract

This paper aimed to discover the risks associated with the dark web and to detect the threats related to human trafficking using image processing with OpenCV and Python. Apart from that, a development environment was set up by installing TensorFlow, OpenCV and Python. Through exploratory data analysis (EDA), significant insights into the distribution and interactions of dataset features were obtained, which are crucial for evaluating various cyberthreats. The construction and evaluation of logistic regression and support vector machine (SVM) models revealed that the SVM model outperforms logistic regression in accuracy. The paper delves into the intricacies of data preprocessing, EDA, and model development, offering valuable insights into network protection and cyberthreat response.