Search Query Log Statistics Generation with Privacy Partitioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data aggregation techniques are incapable of deriving reliable statistical information from large search engine query logs instantaneously, and they pose risks to user privacy by potentially disclosing information about individuals or small groups of users.
Innovation Solution
A computer-implemented method that calculates statistics from user-identified events, selects subsets of database events, merges statistics, and generates results while employing query log sampling, session extraction, and partitioning to ensure privacy and efficiency, using techniques like uniform sub-sampling and tokenization to process query session records.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional data aggregation techniques are used to process search engine query logs, then statistical information can be derived, but the processing time is too long and cannot provide substantially instantaneous responses
Solution Approach 1:
The patent pre-processes and partitions the query log data into multiple subsets before statistical analysis is needed. This preliminary organization allows for much faster querying when statistical information is required, as the system can work with pre-organized data segments rather than processing the entire log from scratch each time.
Solution Approach 2:
The patent divides the large query log database into multiple partitions or subsets. This segmentation allows the system to process smaller data portions in parallel or sequentially, significantly reducing the time required to generate statistical information while maintaining accuracy through proper aggregation of results from all partitions.
2Loss of information
If statistical information is returned based on very small numbers of users or transactions, then specific patterns can be identified, but user privacy is compromised by potentially disclosing information that can be traced back to individuals
Solution Approach 1:
The patent modifies the parameter of data aggregation by grouping transactions into larger clusters or partitions before analysis. This changes the granularity at which statistics are computed, ensuring that individual user patterns are diluted within larger groups, making it infeasible to trace statistics back to specific individuals while still preserving meaningful aggregate patterns.
Solution Approach 2:
The patent introduces an intermediary layer of data aggregation and partitioning between the raw query logs and the statistical output. This intermediary processing ensures that individual user information is transformed and mixed with other users' data before any statistical analysis, acting as a privacy-preserving buffer that prevents direct tracing while maintaining analytical utility.
Data Source
AI summary
A computer-implemented method includes calculating first statistics about a user-identified event within a first subset of a database of events; selecting a second subset of the database of events based on said first statistics; calculating second statistics about the user-identified event within the second subset of the database of events; merging the first and second statistics as statistics of the user-identified event within the entire database of events; and generating a result including at least a portion of the merged statistics of the user-identified event.