Source-linked AI summary
On Facebook, most ties are weak
Pasquale De Meo, Emilio Ferrara, Giacomo Fiumara, Alessandro Provetti
TL;DR
The paper addresses how to classify and evaluate tie strength in online social networks when user-activity data may be unavailable or private. It proposes a topology-only, community-based weak/strong classification and tests it on Facebook and random graphs. The findings suggest weak ties are widespread, robustly identified, and important for information diffusion.
Problem
Existing approaches to tie strength often use user-generated interaction data that may not be publicly available because of privacy concerns.
Method
The paper labels cross-community edges as weak and within-community edges as strong using network topology, then evaluates the definition with community detection and diffusion experiments.
Results
The experiments suggest that weak ties generally outnumber strong ties, the classification is robust across Louvain and Infomap, and weak ties support information spread.
Takeaways & Limitations
Community-based weak ties provide a topology-only way to study cohesion and information diffusion in online social networks.
Abstract
from arXiv · showhide
Pervasive socio-technical networks bring new conceptual and technological challenges to developers and users alike. A central research theme is evaluation of the intensity of relations linking users and how they facilitate communication and the spread of information. These aspects of human relationships have been studied extensively in the social sciences under the framework of the "strength of weak ties" theory proposed by Mark Granovetter.13 Some research has considered whether that theory can be extended to online social networks like Facebook, suggesting interaction data can be used to predict the strength of ties. The approaches being used require handling user-generated data that is often not publicly available due to privacy concerns. Here, we propose an alternative definition of weak and strong ties that requires knowledge of only the topology of the social network (such as who is a friend of whom on Facebook), relying on the fact that online social networks, or OSNs, tend to fragment into communities. We thus suggest classifying as weak ties those edges linking individuals belonging to different communities and strong ties as those connecting users in the same community. We tested this definition on a large network representing part of the Facebook social graph and studied how weak and strong ties affect the information-diffusion process. Our findings suggest individuals in OSNs self-organize to create well-connected communities, while weak ties yield cohesion and optimize the coverage of information spread.
Weak and Strong Ties
The paper adapts weak-tie theory to online social networks by classifying cross-community edges as weak and within-community edges as strong. This topology-based definition avoids restrictive bridge criteria and threshold tuning, but depends on accurate community detection.
- Weak and Strong Ties: Granovetter’s tie strength combines contact time, emotional intensity, intimacy, and reciprocal services.
- Weak and Strong Ties: In online social networks, the classic bridge definition is restrictive because short paths and hubs make complete disconnection unlikely.
- Benefits of our definition: The proposed weak-tie definition labels links between different communities as weak without requiring them to be bridges.
- Benefits of our definition: The classification uses global network topology, avoiding reliance on private user activity or local-neighborhood information.
- Benefits of our definition: Because every edge is labeled weak or strong, the method avoids threshold tuning but cannot compare ties by relative strength.
- Benefits of our definition: The approach depends on accurate community discovery, yet experiments report robustness to the selected community-detection algorithm.
Related Approaches
Related approaches infer tie strength from interaction or other relational variables, often requiring proprietary activity data. The paper instead emphasizes topology and binary community-based classification.
- Related Approaches: Non-topological approaches infer weak ties from measurable relational variables such as interaction frequency, reciprocity, and mutual friends.
- Related Approaches: Facebook research has used private messages, public comments, shared photos, and joint comments to measure connection strength.
- Related Approaches: Unlike the community-based approach, these methods require proprietary user-activity records and assign tie scores classified using thresholds.
Results
Experiments evaluated the definition on a large Facebook sample and random graphs using two community-detection algorithms and information-diffusion simulations. The results indicate algorithmic robustness, widespread weak ties, and an important diffusion role.
- Results: The experiments used Louvain and Infomap on a Facebook sample containing 957,000 users and 58.4 million friendship connections, plus Erdős–Rényi graphs.
- Results: NMI around 0.9 indicates that the communities found by Louvain and Infomap largely coincide.
- Distribution of strong and weak ties in Facebook: At k = 4, weak ties quickly outnumber strong ties among higher-degree Facebook vertices.
- Weak and Strong Ties distribution in Random Graphs: In random graphs, Ravg is always greater than 0.6, so weak ties outnumber strong ties even without a clear community structure.
- The Role of Weak Ties in Information Diffusion: The information-diffusion experiment applied the Independent Cascade Model to Facebook and Erdős–Rényi networks.
- The Role of Weak Ties in Information Diffusion: The simulations selected random starting vertices 250 times and measured diffusion coverage under infection probabilities of 0.01, 0.02, and 0.03.
Conclusions
The paper presents a community-based definition of weak ties for OSNs and concludes that it better fits Facebook and similarly large, dense networks.
- The proposed weak-tie definition is based on the network’s community structure.
- Experiments on a large Facebook sample and randomly generated graphs characterized weak-tie distributions by community size and density.
- The experiments studied weak ties’ role in information-diffusion processes and suggested a connection with Granovetter’s original intuition.
- The authors argue that this community-based definition better fits Facebook and similarly large, dense OSNs than prior approaches.