Muthukumarasamy, Sundar Ayyappan (2024) Automated Detection of Dark Patterns in Website Design: Enhancing User Trust and Online Transparency. Masters thesis, Dublin, National College of Ireland.
Preview |
PDF (Master of Science)
Download (567kB) | Preview |
Preview |
PDF (Configuration Manual)
Download (576kB) | Preview |
Abstract
This project introduces a robust system for the automated detection of dark patterns on websites, aiming to enhance user protection and transparency in online interactions. Leveraging a Naive Bayes classifier trained on dark pattern categories such as Bait and Switch, Forced Continuity, Price Comparison Prevention, Hidden Costs, and Sneaking, the model achieves effective identification of deceptive design elements. The preprocessing of textual data involves employing the TFIDF vectorizer for feature extraction, optimizing the classifier's performance. Web scraping is facilitated through cloud scraping techniques and Beautiful Soup, enabling the extraction of relevant data for classification. The resulting model file is applied to classify scraped data, empowering users to make informed decisions while navigating online interfaces. This innovative approach addresses the ethical concerns associated with dark patterns and contributes to a safer and more transparent online environment.
Actions (login required)
![]() |
View Item |