Host Navigation Graph Clique Mining for Low-Cardinality Host Authority
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques struggle to accurately and efficiently determine the authority of low-cardinality hosts, leading to incomplete analysis and potential misinformation from low-authority sources, while ignoring the importance of low-cardinality hosts in web navigation.
Innovation Solution
A Host Navigation Graph (HNG) is generated using crowd-sourced navigation data, employing the Bron-Kerbosch Clique Graph Mining algorithm to identify cliques and determine host authority values, which are then used to generate GUI elements or mitigation actions to deter interaction with low-authority hosts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing techniques are used to determine host authority, then high-cardinality hosts can be analyzed, but low-cardinality hosts are ignored leading to incomplete analysis
Solution Approach 1:
The patent segments the host analysis problem by separately identifying and analyzing low-cardinality hosts through clique detection in the HNG, rather than treating all hosts uniformly. This segmentation allows the system to focus computational resources on previously overlooked low-cardinality hosts while maintaining analysis of high-cardinality hosts.
Solution Approach 2:
The patent introduces a new dimension of analysis by utilizing the HNG structure and clique relationships to evaluate low-cardinality hosts. Instead of relying solely on traditional metrics that favor high-cardinality hosts, the system adds the dimension of navigation pattern analysis through clique membership to comprehensively assess host authority.
2Measurement precision
If comprehensive host analysis is performed including low-cardinality hosts, then host authority determination accuracy improves, but computational complexity increases
Solution Approach 1:
The patent performs preliminary actions by pre-generating the HNG from crowd-sourced navigation data and pre-identifying cliques within the graph before conducting host authority assessment. This preliminary structuring of data enables efficient querying and analysis of low-cardinality hosts without requiring complex real-time computations during the actual authority determination process.
Solution Approach 2:
The patent creates a simplified representation of host relationships through the HNG and clique structures, which serve as a copied model of the complex web navigation patterns. This copied structure allows for efficient analysis of host authority without directly processing the full complexity of all web navigation data.
3Productivity
If low-cardinality hosts are ignored to reduce computational overhead, then processing efficiency improves, but reliability of host authority assessment deteriorates
Solution Approach 1:
The patent extracts low-cardinality hosts from the general host population by detecting cliques in the HNG, separating them for specialized analysis. This extraction allows the system to maintain high processing efficiency for the majority of hosts while applying targeted, thorough analysis to low-cardinality hosts that would otherwise be overlooked.
Solution Approach 2:
The patent changes the assessment parameters for low-cardinality hosts by utilizing clique membership metrics and HNG-based features rather than traditional authority metrics. This parameter change enables reliable assessment of low-cardinality hosts using metrics specifically suited to their characteristics, improving overall assessment reliability without compromising efficiency.
Data Source
AI summary
Various embodiments of the technology described herein programmatically expand the capabilities of computing systems to allow for the determination of a host authority value indicative of a reliability of a particular host identified as a node in a Host Navigation Graph (HNG). A classifier or other neural network model is trained, based on the HNG, to classify or determine a reliability the particular host based on the host authority value. Based on the host authority value, the particular host is classified and a mitigation is taken to reduce the negative effects associated with a user interacting with a low-authority host. An example mitigation action includes generating a graphical user interface (GUI) element to notify a user of the reliability of a particular host, for example, before, during, or after a user interacts with the host.