Privacy-Preserving Web Recommendation via Anonymized Log Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face information overload when searching online, and current recommendation systems compromise user privacy by using explicit content information from past searches.
Innovation Solution
A method and system that collect raw user log data, transform it to remove user-specific information, and use a machine learning model to generate web page recommendations based on general pages visited, preserving user privacy by leveraging meta-information and collaborative filtering without content data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If recommendation systems use explicit content information from past searches, then recommendation accuracy is improved, but user privacy is compromised
Solution Approach 1:
The patent extracts and removes user-specific information from the log data through anonymization and aggregation processes. Instead of using individual user search queries and content, the system extracts only general page visit patterns that are stripped of personal identifiers, thereby maintaining recommendation accuracy while eliminating privacy risks associated with storing and processing explicit user content data
Solution Approach 2:
The patent introduces an intermediary layer between raw user log data and recommendation generation. This intermediary processing layer includes anonymization, aggregation, and feature extraction steps that transform personal data into generalized patterns. The intermediary prevents direct access to user privacy-sensitive information while preserving the structural patterns needed for accurate recommendations
2Adaptability or versatility
If the system collects and stores detailed user search data, then recommendation personalization is improved, but data privacy risks increase
Solution Approach 1:
The patent fundamentally changes the parameters of data storage and processing by transforming detailed user search data into aggregated statistical patterns. Instead of storing individual user queries, the system stores generalized page visit frequencies and patterns across the user base, maintaining personalization capability through statistical significance while eliminating privacy risks associated with storing detailed personal data
3Measurement precision
If the system processes and stores detailed user interaction data, then recommendation relevance is improved, but system complexity increases
Solution Approach 1:
The patent extracts only the essential patterns from complex user interaction data by aggregating and generalizing page visit information. This extraction process removes unnecessary detailed information while preserving the core patterns needed for relevance, thereby reducing data storage requirements and system complexity while maintaining recommendation quality
Data Source
AI summary
A computer-implemented method of recommending information sources is provided. The method comprising collecting raw user log data of a browser user and transforming the raw user log data to remove data that is specific to the user, wherein the transformed user log data comprises only general pages visited. A number of key page mappings are appended to the transformed user log data. A machine learning model determines a user interaction function according to the transformed user log data and generates a number of web page recommendations, wherein each web page recommendation has a respective probability of engagement based on the user interaction function. The web page recommendations are then displayed to the user on an interface.