Hash-Based Unique Audience Querying via Representation Merging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for determining the number of unique users engaged with multiple pieces of media are resource-intensive, inaccurate, or unable to provide real-time results due to deduplication challenges and precomputing limitations.
Innovation Solution
A system and process that uses hash functions to generate representations of user IDs, allowing for efficient querying of unique audience members across multiple groups by merging and deduplicating these representations, enabling real-time computation without the need for extensive data processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all user identifiers are collected and deduplicated to determine unique users, then accurate unique user count is achieved, but storage space and memory requirements increase significantly
Solution Approach 1:
The patent extracts only the necessary information (user identifiers) from the complete user data and processes only these identifiers for deduplication, rather than storing and processing complete user records. This extraction approach reduces storage requirements while maintaining accurate unique user counting capability.
Solution Approach 2:
The patent segments the user identifier data into manageable portions by organizing data by media item and time period, allowing selective processing of only the segments needed for specific queries rather than processing all data at once, thereby reducing memory requirements.
2Measurement precision
If all user identifiers are collected and deduplicated to determine unique users, then accurate unique user count is achieved, but computation time increases significantly
Solution Approach 1:
The patent performs preliminary actions by pre-computing and storing unique user counts for individual media items and time periods in advance. This allows rapid query responses by simply retrieving pre-computed values rather than performing time-consuming deduplication operations during query time.
Solution Approach 2:
The patent implements dynamic computation strategies where the system adjusts its processing approach based on query characteristics, using pre-computed results for simple queries and performing selective deduplication only when necessary, thereby reducing overall computation time.
3Loss of time
If precomputed unique user counts for individual media items are stored, then query response time improves, but accuracy deteriorates due to double-counting users who engaged with multiple items
Solution Approach 1:
The patent merges pre-computed unique user count data from multiple media items while using set union operations to automatically handle overlaps. This combining approach maintains accuracy by ensuring users are counted only once even though they may have engaged with multiple items, while still enabling fast query responses.
Solution Approach 2:
The patent introduces an intermediary data structure (such as a unified set or temporary storage area) that temporarily holds user identifiers from multiple media items before performing deduplication. This intermediary mechanism allows efficient combination of pre-computed results while maintaining accuracy through systematic overlap resolution.
4Measurement precision
If precomputed values for combinations of media items are stored, then query accuracy improves, but resource consumption increases significantly
Solution Approach 1:
The patent segments the computation into two stages: (1) pre-computing unique user counts for individual media items (low resource consumption), and (2) performing selective combination only when queries require it. This segmentation avoids the resource-intensive approach of pre-computing all possible combinations while maintaining query accuracy.
Solution Approach 2:
The patent applies partial computation by only combining pre-computed results when actually needed for a query, rather than maintaining all possible combination results in advance. This partial action approach reduces storage and computational resource requirements while providing accurate results for actual queries.
Data Source
AI summary
Efficient uniques querying is disclosed, including: receiving a search query for a number of unique audience members across a plurality of groups of audience members; obtaining a plurality of sets of representations corresponding to respective ones of the plurality of groups of audience members; selecting at least a subset from each of the plurality of sets of representations; merging the selected at least subsets of the plurality of sets of representations into a merged set of representations; determining the number of unique audience members across the plurality of groups of audience members based at least in part on the merged set of representations; and outputting the number of unique audience members across the plurality of groups of audience members.