🔒Scroll Patterns Reveal Human vs Bot Traffic
New metrics can spot bots on your site
TL;DR
Researchers use Burstiness (B) and Memory (M) to distinguish between human users and bots based on scroll behavior. The LightGBM model achieved 73.4% accuracy but struggles with ChatGPT Agent.
Scientists have developed a method using Burstiness (B) and Memory (M) metrics to identify whether website traffic is from humans or bots, focusing on scroll patterns. This approach leverages the distinct bursty nature of human scrolling versus the more uniform behavior of bots. The model trained on these metrics achieved an accuracy of 73.4% in distinguishing between humans and ChatGPT Agent sessions. However, its effectiveness may be limited on sites that don't require much scrolling or have different interaction patterns. Why does this matter? Website owners can use these insights to better protect their content from scraping bots by combining scroll behavior analysis with other features like mouse movements or typing patterns. This could significantly enhance security measures for websites dealing with sensitive data or high traffic.
Key Points
Burstiness (B) and Memory (M) metrics were calculated for JavaScript 'scroll' events per page visit to distinguish between humans and bots.
A LightGBM model trained on these metrics achieved an accuracy of 73.4% in classifying human vs bot sessions based on scroll patterns.
Human distribution showed a higher Burstiness coefficient (1.2) and almost no Memory, while ChatGPT Agent had similar B and M values to humans.
Scroll behavior analysis may not be effective on websites that don't require scrolling or have different interaction patterns.
Combining this method with other features like mouse movements or typing could create a more robust model for detecting bots.
Why It Matters
Website owners can use scroll pattern analysis to better protect their content from scraping bots. For instance, combining bursty scroll detection with mouse movement tracking on high-traffic sites could significantly enhance security measures against automated scrapers.
Frequently Asked Questions
Why does this matter?
Website owners can use scroll pattern analysis to better protect their content from scraping bots. For instance, combining bursty scroll detection with mouse movement tracking on high-traffic sites could significantly enhance security measures against automated scrapers.
What happened?
Researchers use Burstiness (B) and Memory (M) to distinguish between human users and bots based on scroll behavior. The LightGBM model achieved 73.4% accuracy but struggles with ChatGPT Agent.
Comments
Be the first to comment
Enjoyed this article?
Get it daily. 7am. Free. Reads in 5 minutes.
Join 3,265 builders reading daily.