Predicting User Interests Based on Who Looks at Whose Profile

A method for predicting what a user might be interested in by analyzing the web of connections created when people view each other's social media profile pages.

What it covers

This patent describes a way to map social network activity by tracking who visits whose profile page. It creates a graph where nodes represent users and directional links represent a user viewing another user's profile. If a user is known to like a specific topic, the system assigns them a score and then propagates that score to other users who have viewed the same profiles. This allows the system to calculate a likelihood score for other users, suggesting they might also be interested in that same topic based on their browsing behavior.

What it doesn't cover

• —Does not cover analyzing content based on text keywords or hashtags within a user's own posts.
• —Does not cover predicting interests based on direct social connections like 'friends' or 'followers'.
• —Does not cover real-time tracking of mouse movements or dwell time on a specific image.
• —Does not cover interest prediction based on external search engine queries.

The clever bit

The system treats the act of 'viewing a profile' as a signal of shared interest, effectively using the social graph as a proxy for taste, even if the users never interact directly.

Why it matters

This technology is a foundation for interest-based recommendation engines. By using behavioral data—who looks at what—rather than just profile data, companies can build more accurate models of user preferences. It was filed during the early growth of social networks, when platforms were moving from simple directories to complex, data-driven advertising engines.

Real-world examples

• 1.Targeted advertising on social media platforms
• 2.Content recommendation feeds
• 3.Suggested user or interest discovery features