Short Identifier Analytics With Collision-Adjusted Behavioral Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The use of short identifiers in behavioral analytics can result in identifier collisions, leading to less accurate pattern recognition and future predictions, especially when the number of entities exceeds the maximum unique identifier capacity, such as with short integers with a maximum of four digits.
Innovation Solution
A computing system and method that determines identifier collisions between geographically dispersed locations, performs behavioral analytics on short identifiers, generates adjusted behavioral models, and notifies users of incremental event values based on these models, using collision decay and cohort analysis to mitigate the effects of identifier collisions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If short identifiers are used to reduce storage requirements and increase processing speeds, then storage efficiency and processing speed are improved, but identifier collisions occur leading to reduced accuracy in pattern recognition
Solution Approach 1:
The patent introduces an intermediary mechanism (collision detection and adjustment module) that mediates between the short identifier system and the behavioral analytics processing. This intermediary identifies collisions and adjusts the analysis results accordingly, allowing the system to maintain the speed benefits of short identifiers while correcting for the accuracy losses they cause.
Solution Approach 2:
The system implements feedback by continuously monitoring for identifier collisions and using this information to adjust behavioral models. The collision detection feedback loop allows the system to identify when short identifiers have caused misattribution and correct the analytical results, maintaining accuracy despite using compressed identifiers.
2Quantity of substance
If short identifiers are used to reduce storage requirements, then storage efficiency is improved, but identifier collisions occur leading to less accurate future predictions
Solution Approach 1:
An intermediary collision adjustment module processes prediction results to correct for identifier collisions. This mediator allows the system to store and process data using space-efficient short identifiers while maintaining prediction accuracy by adjusting results based on detected collision patterns.
Solution Approach 2:
The system converts the harmful effect of identifier collisions into beneficial information. By detecting where collisions occur, the system learns about data distribution patterns and uses this knowledge to improve future predictions, turning the limitation of short identifiers into a source of analytical insight.
3Adaptability or versatility
If the number of entities exceeds the maximum unique identifier capacity, then the system can handle more entities, but identifier collisions become inevitable
Solution Approach 1:
The patent introduces a collision detection and adjustment intermediary that enables the system to handle more entities than the identifier space would normally allow. This mediator identifies when multiple entities share the same short identifier and adjusts the behavioral analysis to account for this ambiguity, maintaining reliability despite exceeding identifier capacity.
Solution Approach 2:
The system performs partial identification using short identifiers while accepting that complete uniqueness cannot be achieved. Instead of requiring full identifier uniqueness, the system uses short identifiers for efficient processing and applies statistical adjustments to compensate for the inevitable collisions, enabling handling of excessive numbers of entities.
Data Source
AI summary
Computing systems, methods, and non-transitory computer-readable storage media are provided. Identifier collisions are determined in historical event data based on short identifiers associated with geographically dispersed first and second event locations. Behavior analytics are performed on multiple short identifiers in the historical event data to generate behavioral models associated with the multiple short identifiers. Adjusted behavioral models are generated based on the determined identifier collisions. A short identifier is obtained and an adjusted behavioral model is determined. A new event value for a first event type is calculated when a total value of events of the first event type during a current analysis period does not exceed a baseline value associated with the first event type. A client device is notified using the new event value, whereby the client device displays a notification regarding the new event value in response to receiving the new event value.


