Seed Node Selection Using Metadata in Unknown Network Topology
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing research on influence maximization in social networks is limited by the lack of complete connection information, making it difficult to select seed nodes that maximize influence when only limited edge information is available.
Innovation Solution
A network node selection method that estimates edge probabilities using node metadata and expands the network by selecting query nodes based on these probabilities, creating a virtual edge graph and applying diffusion models to identify high-influence seed nodes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If query transmission method is used to obtain edge information, then network connection information can be expanded, but the number of queries is limited and each query incurs a fee
Solution Approach 1:
The patent applies preliminary action by pre-calculating virtual edge probabilities for all possible node pairs using node metadata before actually transmitting queries. This allows the system to prioritize which queries to transmit based on pre-computed likelihoods, maximizing information gain while minimizing the number of expensive query transmissions needed to expand the network connection information.
Solution Approach 2:
The patent creates virtual edges as probabilistic copies of potential real edges based on node metadata similarities. These virtual edges serve as approximations that guide query selection, allowing the system to infer likely connections without transmitting queries for every possible node pair, thus expanding information with fewer actual queries.
2Productivity
If seed node selection is performed with limited edge information, then influence maximization can be achieved, but the accuracy of seed node selection deteriorates without complete topological structure
Solution Approach 1:
The patent introduces virtual edge probabilities as an intermediary mechanism that bridges the gap between limited observed edge information and the complete topological structure needed for accurate seed node selection. These probabilistic intermediaries allow the influence maximization algorithm to operate with incomplete information while maintaining selection accuracy by accounting for potential unseen connections.
Solution Approach 2:
The patent transforms the seed node selection problem from a discrete binary decision (edge exists or not) to a continuous probabilistic framework by assigning virtual edge probabilities. This parameter change allows the algorithm to consider the likelihood of edge existence, enabling more accurate influence maximization even when the actual edge structure is partially unknown.
3Loss of information
If all possible virtual edges are created between nodes, then the network expansion is comprehensive, but the computational complexity increases
Solution Approach 1:
The patent reduces computational complexity by transforming the exhaustive edge enumeration problem into a probabilistic scoring problem. Instead of treating all virtual edges equally, the system computes probability scores based on node metadata and uses these scores to prioritize which virtual edges to explore through actual queries, making the comprehensive network expansion computationally tractable.
Solution Approach 2:
The patent applies partial action by creating and evaluating virtual edges for all node pairs but only pursuing actual queries for the most promising candidates based on probability thresholds. This allows the system to maintain comprehensive network coverage in principle while practically focusing computational resources on the most likely beneficial connections.
Data Source
AI summary
A network node selection apparatus selects a seed node with the greatest influence by expanding the network connection structure by repeating, for a limited number of queries, the process of estimating an edge probability indicating the possibility of the existence of an edge between nodes based on the metadata of a plurality of nodes separately collected in a network where connection information between nodes through the edge is not provided, and expanding the network by selecting a query node according to the estimated edge probability, and selecting a seed node from the expanded network.