Related Experiment Videos
Patterns of link reciprocity in directed networks.
Diego Garlaschelli1, Maria I Loffredo
1Dipartimento di Fisica, Università di Siena, Via Roma 56, 53100 Siena, Italy.
Physical Review Letters
|February 9, 2005
Summary
This study introduces a new network analysis measure for link reciprocity, finding real-world networks exhibit consistent correlation patterns. These findings challenge existing models, proposing a new framework for understanding mutual links.
Area of Science:
- Network Science
- Graph Theory
- Data Analysis
Background:
- Link reciprocity, the presence of mutual links between network nodes, is a fundamental property.
- Existing network models fail to capture the observed correlation patterns in real-world networks.
Purpose of the Study:
- To develop a novel measure for quantifying link reciprocity.
- To analyze and compare reciprocity across diverse network types.
- To propose an improved framework for modeling mutual links.
Main Methods:
- Development of a new reciprocity measure for network analysis.
- Empirical analysis of various real-world networks (economic, social, cellular, etc.).
- Introduction of a conditional connection probability framework.
Main Results:
- Real-world networks consistently show either correlated or anticorrelated link reciprocity.
- Networks of the same type exhibit similar reciprocity values.
- Current network models do not replicate these observed reciprocity patterns.
Conclusions:
- The proposed reciprocity measure effectively orders networks by mutual link correlation.
- A new framework incorporating conditional connection probability is needed to explain observed patterns.
- Reciprocity patterns are type-dependent and not universally random.