Social Influence Detection via Data Segmentation and Intermediary Tests
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Detecting social influence in social networks is challenging due to its similarity with other factors like homophily and confounding variables, making it difficult to distinguish and quantify.
Innovation Solution
The method involves analyzing social network data by modifying relationships and timestamps using tests like the shuffle test and edge reversal test to compare initial and modified data estimates, determining social influence through logistic regression and coefficient analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional correlation analysis is used to detect social influence, then measurement capability is provided, but measurement precision deteriorates due to inability to distinguish social influence from homophily and confounding variables
Solution Approach 1:
The patent segments the correlation analysis into multiple distinct components by creating different test versions (original, time-shuffled, edge-shuffled, reversed) that isolate specific factors. Each test version measures a different aspect of social correlation, allowing the system to separate social influence effects from homophily and confounding variables through comparative analysis of these segmented measurements
Solution Approach 2:
The patent introduces shuffled and reversed test versions as intermediary comparisons to mediate between the original correlation measurement and the underlying social influence effect. These intermediary tests act as control groups that capture confounding factors and homophily, allowing the true social influence to be extracted by comparing against these intermediary measurements
2Measurement precision
If data modification tests are applied to distinguish social influence, then measurement precision improves, but device complexity increases due to multiple processing steps
Solution Approach 1:
The patent creates multiple copies of the social network data with different transformations applied (time-shuffled copy, edge-shuffled copy, reversed copy). These copied datasets serve as control groups that preserve the overall structure while randomizing specific factors, enabling precise isolation of social influence effects through comparison without requiring complex real-time processing
Solution Approach 2:
The patent systematically changes specific parameters of the social network data (time stamps, edge directions, association relationships) while keeping other structures intact. By changing only one parameter at a time in different test versions, the system can isolate the effect of each parameter change and attribute differences in correlation measurements to specific factors like social influence versus homophily
Data Source
AI summary
Techniques for detecting social influence between users in a set of users with regard to an activity are provided. Data for each user of the set of users is received that includes a time value at which the user became active with regard to the activity, and includes at least one indication of another user in the set of users associated with the user. A first estimate of social correlation in the set of users is determined based on the data. The data is modified. For instance, the data may be modified according to a shuffle test and/or an edge reversal test. A second estimate of social correlation in the set of users is determined based on the modified data. The first estimate is compared to the second estimate to determine a degree of social influence in the set of users.