URL Set Event Detection for Targeted Network Response
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Machine learning models struggle to effectively learn from vast and complex URL histories, making it difficult to identify and respond to events driven by increased accessing of uniform resource locators (URLs) in communication networks.
Innovation Solution
A processing system tracks URL accessing patterns to identify events, applies actions to a test group of users associated with these events, and adjusts network resources based on success rates to address demands, using techniques like re-routing, load-balancing, and denial-of-service mitigation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine learning models are used to learn from URL histories, then event detection capability is improved, but the complexity of processing voluminous and complex data increases
Solution Approach 1:
The patent segments the complex URL history data by grouping URLs into sets based on their relationships (e.g., same domain, similar content). This segmentation reduces the complexity of processing individual URLs while maintaining the ability to detect events through aggregated analysis of URL sets.
Solution Approach 2:
The patent introduces URL sets as an intermediary layer between raw URL data and event detection. Instead of directly analyzing individual URLs, the system analyzes patterns within URL sets, which serves as a mediator that simplifies the detection process while preserving event identification capability.
2Reliability
If actions are applied to the entire user group, then event response coverage is improved, but the risk of applying ineffective actions increases
Solution Approach 1:
The patent applies actions to only a test group (a portion) of users associated with an event rather than the entire user group. This partial action approach allows the system to test effectiveness with a subset of users before considering broader application, reducing the waste of resources on ineffective actions.
Solution Approach 2:
The patent implements a feedback mechanism by monitoring the success rate of actions applied to the test group. Based on this feedback, the system determines whether to expand the action to additional portions of the user group, ensuring that actions are only broadly applied when proven effective.
3Measurement precision
If URL accessing patterns are tracked for all users, then event identification accuracy is improved, but the data volume and processing burden increase
Solution Approach 1:
The patent merges individual URL accessing patterns into aggregated URL set patterns. By combining data from multiple users and grouping related URLs together, the system maintains high event identification accuracy through pattern recognition while reducing the overall data volume that needs to be processed individually.
Data Source
AI summary
A processing system may track accessing of a plurality of uniform resource locators among users of a communication network, determine an occurrence of an event of an event type based upon an increased accessing of the plurality of uniform resource locators for at least one time period, identify a first sub-group of the users associated with the event, apply a first action in the communication network to a test group including at least a first portion of the first sub-group, the first action addressing a demand associated with the event of the event type, track at least a first success rate of the first action for the first portion of the first sub-group, and apply the first action to at least a second portion of the first sub-group in response to determining that the first success rate for the first portion of the first sub-group exceeds a threshold success rate.