User Activity Attribution in Multi-User Machine Settings
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing online services face challenges in accurately attributing user activity to individual users on shared machines, leading to noisy signals for personalization and advertising, as shared machines complicate the mapping of user identifiers to individual users.
Innovation Solution
A system and method that train a classifier to distinguish between single-user and multi-user machines based on activity logs, estimate the number of users, cluster user activity, and assign new activity using a similarity function to accurately attribute search behavior to specific users.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If machine identifiers are used to track user activity on shared machines, then online services can provide personalized offerings, but the accuracy of user attribution deteriorates due to interwoven search histories from multiple users
Solution Approach 1:
The patent segments the aggregated machine activity into distinct user-specific activity patterns by analyzing temporal, behavioral, and contextual features. The system divides the mixed search history into separate user profiles through clustering algorithms that identify distinct usage patterns, thereby resolving the contradiction between maintaining personalization and achieving accurate user attribution.
Solution Approach 2:
The patent introduces an intermediary layer of user identification that sits between the machine identifier and the online service. This intermediary uses activity logs and behavioral analysis to infer individual user identities without requiring direct user identification, enabling accurate attribution while maintaining personalization capabilities.
2Quantity of substance
If activity logs from multiple users are aggregated under a single machine identifier, then comprehensive usage data is collected, but the signal quality deteriorates due to noisy and interwoven user behaviors
Solution Approach 1:
The patent extracts distinct user-specific signals from the aggregated machine activity logs by identifying and separating characteristic behavioral patterns. The system takes out individual user activity patterns from the mixed data stream through feature analysis and clustering, preserving signal quality while maintaining comprehensive data collection.
Solution Approach 2:
The patent applies local quality by treating different segments of the activity log with different analytical approaches. The system identifies regions or time periods dominated by specific users and applies user-specific analysis to those segments, thereby preserving the quality of individual user signals while maintaining the comprehensive nature of the aggregated data.
Data Source
AI summary
The claimed subject matter includes a system and method for attribution of search activity in multi-user settings. The method includes training a classifier to distinguish between machines that are single-user and multi-user based on activity logs of an identified machine. The identified machine is determined to be multi-user based on the classifier. A number of users is estimated for the identified machine. Activity of the users is clustered based on the number of users estimated. A similarity function is learned for the number of users estimated. The method also includes assigning new activity to one of the users based on the clustering, and the similarity function.