A comparative study of machine learning: Models for web tracker detection

Title

A comparative study of machine learning: Models for web tracker detection

Creator

Mathew, Calvin; Poonia, Ramesh Chandra

Description

Web trackers, used by websites, collect user data and monitor online activity, often with or without explicit consent. With concerns for online privacy, there is a growing need to detect these web trackers. This study evaluates several machine learning (ML) techniques for detecting web trackers, focusing on evaluating their performance from the key metrics such as accuracy, precision, and recall. We analyzed supervised methods, such as random forest, support vector machines (SVM), neural networks, gradient boosting, and unsupervised methods, including DBSCAN and isolation forest. Models were trained on a comprehensive dataset extracted from URLs with feature engineering, and data preprocessing techniques were applied to enhance model performance and detect both known and unknown trackers and normal traffic. Our results indicate that supervised models outperform unsupervised methods, demonstrating their superior ability in distinguishing web trackers from normal traffic. This work highlights the effectiveness of ML-based tracker detection and outlines opportunities for improving privacy protection through adaptive supervised learning methods. 2026 selection and editorial matter, Jossy George, Kamal Upreti, Ramesh Chandra Poonia, Ankit Gautam, and Danish Nadeem; individual chapters, the contributors.

Source

Cognitive Cloud Computing: Building Intelligent Systems for Tomorrow;pp.245-263

Date

01-01-2025

Publisher

Taylor and Francis

Coverage

Mathew C., School of Sciences, CHRIST University, Delhi NCR, India; Poonia R.C., Department of Computer Science, CHRIST (Deemed to be University, Bengaluru, India

Rights

Restricted Access; Hardcopy may be available in the library

Relation

ISBN: 978-104054278-1; 978-103294165-3;

Format

online

Language

English

Type

Book chapter

Identifier

https://doi.org/10.1201/9781003569251-13

https://www.scopus.com/pages/publications/105024615475?origin=resultslist