[Paper Review] Community Detection in Social Network using Temporal Data

Victor Stany Rozario, AZM Ehtesham Chowdhury+1 more|arXiv (Cornell University)|Apr 10, 2019

Complex Network Analysis Techniques12 references4 citations

TL;DR

This paper proposes the Interlinked Spatial Clustering Model (ILSCM), a novel community detection approach that integrates temporal data, node attributes, and edge connectivity to improve clustering accuracy in social networks. By leveraging context-based key extraction and temporal topic modeling, ILSCM identifies communities through dynamic betweenness analysis, demonstrating enhanced detection performance over traditional methods that ignore temporal dynamics.

ABSTRACT

Social networks facilitate the social space where actors or the users have ties among them. The ties and their patterns are based on their life styles and communication. Similarly, in online social media networks like Facebook, Twitter, Myspace etc., multiple users belong to multiple specific communities. The social network represents large real-world directed social graphs. Detection of communities or clusters from these graphs is a problem of considerable research interest. The communities are formed using the neighboring nodes that have common edges and common attributes. Most of the existing community detection algorithms usually consider node contents to analyze the attributes of community. Some algorithms use the links between the nodes to determine the dense regions in the graph. But utilizing both the edge content and the vertex content to detect community are yet to be considered and verified, since the traditional extraction methods of vertex and edge data do not consider the connectivity among the nodes. This paper presents an Interlinked Spatial Clustering Model (ILSCM) which provides relevant content selection and extraction of the temporal topics for identifying the betweenness among the nodes based on the context keys to detect community.

Motivation & Objective

• To address the limitation of existing community detection algorithms that fail to integrate temporal dynamics with node and edge attributes.
• To develop a method that captures evolving community structures in social networks using time-stamped interaction data.
• To improve community detection accuracy by jointly analyzing vertex content, edge connectivity, and temporal patterns.
• To introduce a context-aware content selection mechanism that enhances relevance in community identification.
• To validate the effectiveness of combining temporal, spatial, and attribute-based features in clustering social network graphs.

Proposed method

• The Interlinked Spatial Clustering Model (ILSCM) is proposed to detect communities by analyzing temporal patterns in user interactions.
• Temporal topic modeling is applied to extract evolving topics from user interactions over time, forming context keys for clustering.
• Betweenness centrality is computed based on context keys to identify key nodes and linkages that define community boundaries.
• Node attributes and edge connections are integrated with temporal data to enrich feature representation for clustering.
• A spatial clustering approach is applied to group nodes based on combined temporal, attribute, and connectivity features.
• The model uses context-based content selection to prioritize relevant interactions and reduce noise in community formation.

Experimental results

Research questions

• RQ1 How can temporal dynamics in social network interactions improve community detection accuracy?
• RQ2 To what extent do combined node attributes and edge connectivity enhance community identification compared to isolated feature use?
• RQ3 Can context-based key extraction from temporal data improve the detection of evolving communities?
• RQ4 How does integrating temporal topics with structural connectivity affect clustering performance?
• RQ5 What is the impact of using both vertex and edge content in a dynamic network setting?

Key findings

• The ILSCM model successfully identifies communities by integrating temporal, attribute, and connectivity data, outperforming methods that use only static or isolated features.
• Temporal topic modeling enables the detection of evolving community structures that are not visible in static network analysis.
• The use of context keys for betweenness analysis improves the identification of pivotal nodes and inter-community links.
• The model demonstrates enhanced clustering accuracy due to the synergistic use of vertex content, edge content, and time-stamped interactions.
• The integration of temporal data reduces noise and increases relevance in community detection, particularly in dynamic social networks.
• Empirical results show that ILSCM achieves better community cohesion and separation compared to baseline algorithms on real-world social network data.