Random-Walk Node Ranking With Edge Sign Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing personalized node ranking (PNR) methods in signed networks fail to accurately consider edge signs, leading to inaccurate rankings due to incorrect propagation of positive and negative scores based on the balance theory.
Innovation Solution
A method and system for personalized node ranking using random walk that verifies score propagation by predicting edge signs through topological features and using a logistic regression classifier model to determine trustworthiness, and adjusting propagation based on balanced and unbalanced triangles in the network.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If balance theory is used to propagate scores in signed networks, then score propagation follows theoretical rules, but false sign predictions cause inaccurate rankings
Solution Approach 1:
The patent performs sign verification before score propagation by training a classification model on topological features to predict edge signs in advance. This preliminary action identifies unreliable edge signs before they can cause false score propagation, thereby resolving the contradiction between following balance theory rules and avoiding inaccurate rankings.
Solution Approach 2:
The patent implements a feedback mechanism where the classification model's prediction results are used to verify and correct potential errors in balance theory-based score propagation. The model provides feedback on whether propagated scores are reliable, allowing the system to adjust propagation behavior based on predicted edge sign accuracy.
2Adaptability or versatility
If existing PNR methods are used in signed networks, then global node ranking is provided, but edge signs are not considered leading to inaccurate results
Solution Approach 1:
The patent applies local quality by treating positive and negative edges differently in the score propagation process. Instead of uniform propagation, the system adjusts propagation behavior based on local edge sign characteristics and their predicted reliability, thereby adapting to the signed network structure while maintaining ranking accuracy.
Solution Approach 2:
The patent changes the propagation parameter (score reliability) based on edge sign predictions. By dynamically adjusting whether to propagate scores along certain edges based on the classification model's output, the system adapts to signed network characteristics while preventing inaccurate propagation.
3Reliability
If sign verification is performed using classification model, then false sign propagation is prevented, but system complexity increases
Solution Approach 1:
The patent introduces a classification model as an intermediary between the balance theory propagation and the final ranking results. This intermediary verifies edge sign reliability without requiring fundamental changes to the propagation mechanism, thereby preventing false propagation while adding manageable complexity through a dedicated verification layer.
Solution Approach 2:
The patent segments the score propagation process into two distinct phases: verification phase (using classification model on topological features) and propagation phase (using verified scores). This segmentation isolates the complexity of sign verification from the propagation mechanism, making the system more manageable and interpretable.
Data Source
AI summary
Provided is a personalized node ranking (PNR) system and method using random walk. In detail, a personalized ranking method using random walk includes verifying whether propagation of a positive score or a negative score of a currently-visiting node is trustworthy based on an edge sign between a seed node and a next-visiting node predicted by using a classification model, and when the propagation is trustworthy according to the verification, propagating a score by using a first score propagation method based on balance theory, and when the propagation is untrustworthy according to the verification, propagating a score by using a second score propagation method based on a ratio between a balanced triangle and an unbalanced triangle of a signed network.