Unique User Count Correction Using Shared Resource Boundaries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for estimating unique resource users across a group of arbitrarily grouped resources in custom properties often result in inaccurate aggregate counts due to errors in sampling rates and user representation, leading to significant deviations from the actual count.
Innovation Solution
A system and method that utilize graph-based composition capping to correct estimated counts of unique resource users by determining composition boundaries and adjusting counts based on minimum and maximum constraints within hierarchical graphs, incorporating bipartite and hierarchical graph structures to account for shared resources and user overlaps.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If individual unique user counts are determined separately for each resource in a group, then the calculation process is simple, but the aggregate count becomes inaccurate due to user overlaps across multiple resources
Solution Approach 1:
The patent merges the determination of unique user counts across multiple resources by introducing a supervisor entity that coordinates between resources. Instead of completely separate calculations, the supervisor collects individual resource counts and overlap information, then combines them using inclusion-exclusion principles to produce an accurate aggregate count that accounts for users appearing in multiple resources.
Solution Approach 2:
The patent introduces a supervisor entity as an intermediary between individual resources. This supervisor receives unique user counts from each resource, determines overlap information between resource user sets, and calculates the corrected aggregate count. The supervisor acts as a mediator that resolves the accuracy issue without requiring complete redesign of the individual resource counting process.
2Adaptability or versatility
If resources are grouped into custom properties with arbitrary groupings, then resource organization flexibility is improved, but the accuracy of aggregate unique user counts deteriorates due to unaccounted user overlaps
Solution Approach 1:
The patent introduces dynamic adjustment mechanisms that allow the system to adapt to different resource grouping configurations. The supervisor entity dynamically determines overlap information based on the specific custom property groupings, and the correction process dynamically adjusts the aggregate count based on the actual overlap patterns. This maintains flexibility in resource organization while ensuring accuracy through adaptive correction.
Solution Approach 2:
The patent implements a feedback mechanism where the supervisor receives information about user overlaps from individual resources, processes this information to determine corrections, and applies adjustments to the aggregate count. The system continuously monitors and adjusts for overlaps, providing feedback-based correction that maintains accuracy regardless of how resources are grouped into custom properties.
3Productivity
If sampling rates are used to estimate unique user counts, then data collection efficiency is improved, but estimation accuracy deteriorates due to sampling errors and user representation biases
Solution Approach 1:
The patent applies feedback principles by using the supervisor to collect sampling information from multiple resources, analyze the overlap patterns, and adjust the aggregate estimation accordingly. The system uses the sampling data feedback to identify and correct for estimation errors, improving the overall accuracy of unique user count calculations while maintaining the efficiency benefits of sampling-based data collection.
Data Source
AI summary
Correction of estimated unique counts of resource users is disclosed, including: determining a set of shared resources between a first property and a second property, wherein the first property comprises a custom property comprising a configurable collection of two or more resources; recording a hidden node that corresponds to the set of shared resources; obtaining an estimated aggregate count of unique resource users corresponding to the first property; obtaining an estimated deduplicated count of unique resource users across the set of shared resources, wherein the estimated deduplicated count of unique resource users corresponds to the hidden node; determining minimum and maximum boundaries corresponding to the first property based on the estimated deduplicated count of unique resource users corresponding to the hidden node; and correcting the estimated aggregate count of unique resource users corresponding to the first property based on the minimum and maximum boundaries.