Predictive Behavioral Centrality for Influencer Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for identifying influencers in social networks rely on structural centrality measures that do not account for specific behaviors and are not scalable for large networks, often misrepresenting influence and requiring unrealistic propagation models.
Innovation Solution
A method that uses behavioral centrality measures derived from past data to predict the most probable influencers by analyzing communication patterns and adoption of specific behaviors within a social network, allowing for targeted selection of entities based on predicted influence and reach.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If structural centrality measures are used to identify influencers, then the selection process is simple and fast, but the accuracy of influencer identification deteriorates because these measures do not account for specific behaviors
Solution Approach 1:
The patent transitions from structural parameters (degree, betweenness, closeness centrality) to behavioral parameters (adoption patterns, propagation effectiveness) to measure influence. This parameter change allows the system to maintain computational efficiency while significantly improving the accuracy of influencer identification by focusing on actual behavioral data rather than just network topology.
Solution Approach 2:
The patent replaces the mechanical calculation of structural centrality measures with a data-driven approach using machine learning models that analyze behavioral patterns. Instead of relying on fixed mathematical formulas based on network structure, the system uses trained models that process historical behavioral data to predict future influence, substituting rigid mechanical computation with adaptive statistical learning.
2Reliability
If propagation models are used to evaluate influencer performance, then theoretical understanding of influence mechanisms is improved, but the complexity of the system increases and scalability deteriorates for large networks
Solution Approach 1:
The patent employs lightweight, computationally inexpensive predictive models that can be rapidly trained and applied to large networks. Instead of using complex, resource-intensive propagation models that require extensive computational power and long processing times, the system uses simplified machine learning models that provide sufficient accuracy while being scalable to millions of nodes, effectively replacing heavy theoretical models with efficient practical alternatives.
3Ease of manufacture
If standard market research techniques are used to define influencers by specific attributes, then the classification process is straightforward, but the ability to predict future influence deteriorates because past attributes do not necessarily indicate future behavior
Solution Approach 1:
The patent performs preliminary training of machine learning models using historical behavioral data before applying them to identify influencers. This preliminary action involves collecting and analyzing past adoption patterns and propagation events to train predictive models, which then capture dynamic behavioral patterns rather than relying on static attribute classification. This prepares the system in advance to accurately predict future influence based on learned patterns rather than predefined categories.
Data Source
AI summary
The method uses predictive analysis to determine a model based on past data including a first social network built between communicating entities for a first observation period and behavioral centrality measures derived from behavioral data observed in a following time period. The model thus determined is then applied to a second social network built for a second observation period more recent than the first one. This provides predicted behavioral centrality measures for a future period, which can be used to perform an efficient selection of entities in the target, which may maximize virality with respect to the specific behavior of interest.