Year: 2025
Type: article
Abstract: In today’s digital landscape, deceptive design patterns—such as fake urgency messages, misleading buttons, and disguised advertisements—are increasingly used to manipulate user behavior on websites. This project, titled “Deceptive Behavior Analysis Using Deep Learning,” presents an automated solution to detect such deceptive elements. It uses a Chrome extension to capture webpage screenshots and DOM content, which is then processed by a Flask backend. Text is extracted using the EAST deep learning model, vectorized using TF-IDF, and analyzed using two machine learning classifiers: Bernoulli Naive Bayes to detect the presence of deception, and Multinomial Naive Bayes to categorize the type of deceptive pattern. The results are stored in Excel and CSV for analysis. This system offers a scalable, real-time approach to identifying deceptive behaviors on websites and enhancing user protection.
Authors: Vijayajyothi Chiluka, Pravalika Bandi, Pranathi Enumula, Laxmi Sowjanya Korvi, Varun Teja Seelam
Language: English
Topic: Spam and Phishing Detection
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W4410241264