[Paper Review] A Survey of Community Search Over Big Graphs

Yixiang Fang, Xin Huang+5 more|arXiv (Cornell University)|Apr 29, 2019

Complex Network Analysis Techniques210 references9 citations

TL;DR

This survey provides a comprehensive review of community search (CS) over large-scale graphs, analyzing cohesiveness metrics like k-core, k-truss, and k-clique, and evaluating CS methods across diverse graph types. It identifies key challenges in efficiency, scalability, and dynamic updates, and proposes future research directions for big, heterogeneous, and uncertain graphs.

ABSTRACT

With the rapid development of information technologies, various big graphs are prevalent in many real applications (e.g., social media and knowledge bases). An important component of these graphs is the network community. Essentially, a community is a group of vertices which are densely connected internally. Community retrieval can be used in many real applications, such as event organization, friend recommendation, and so on. Consequently, how to efficiently find high-quality communities from big graphs is an important research topic in the era of big data. Recently a large group of research works, called community search, have been proposed. They aim to provide efficient solutions for searching high-quality communities from large networks in real-time. Nevertheless, these works focus on different types of graphs and formulate communities in different manners, and thus it is desirable to have a comprehensive review of these works. In this survey, we conduct a thorough review of existing community search works. Moreover, we analyze and compare the quality of communities under their models, and the performance of different solutions. Furthermore, we point out new research directions. This survey does not only help researchers to have a better understanding of existing community search solutions, but also provides practitioners a better judgment on choosing the proper solutions.

Motivation & Objective

• To systematically review and compare community search (CS) techniques for large-scale graphs across diverse graph types and cohesiveness metrics.
• To identify critical challenges in real-time performance, scalability to billion-scale graphs, and support for dynamic or uncertain graph updates.
• To highlight gaps in current research, particularly the lack of CS solutions for heterogeneous, multi-dimensional, and signed graphs.
• To advocate for an open repository of CS code and datasets to accelerate benchmarking and practical deployment.
• To guide researchers and practitioners in selecting appropriate CS methods based on graph characteristics and application requirements.

Proposed method

• Classify CS methods based on cohesiveness metrics such as k-core, k-truss, k-clique, and k-ECC, analyzing their structural and computational properties.
• Survey 30+ studies from 2010–2019, categorizing them by graph type (e.g., static, uncertain, signed, multi-dimensional, HINs) and cohesiveness model.
• Evaluate performance and quality of communities using metrics like connectivity, density, and interpretability across different algorithms.
• Propose distributed and I/O-efficient computation models to scale CS to billion-vertex graphs that exceed single-machine memory.
• Introduce a framework for online CS that supports real-time query processing with personalized constraints on query vertices and structural attributes.
• Highlight the need for standardized benchmarks and public code/dataset repositories to enable reproducibility and practical adoption.

Experimental results

Research questions

• RQ1 How do different cohesiveness metrics (e.g., k-core, k-truss) affect the quality and efficiency of community search in large graphs?
• RQ2 What are the key limitations of existing CS methods in handling real-world graph characteristics such as uncertainty, dynamics, and heterogeneity?
• RQ3 How can CS algorithms be scaled to billion-vertex graphs that cannot fit in main memory?
• RQ4 What are the open challenges in supporting community search over emerging graph types like heterogeneous information networks (HINs) and multi-dimensional graphs?
• RQ5 How can an open repository of CS code and datasets improve reproducibility and practical deployment in industry and research?

Key findings

• Most existing CS methods are limited to static, undirected graphs and do not support dynamic updates, uncertain edges, or complex graph types like HINs.
• Only a few studies (e.g., Fang:TKDE:CSD, Li:vldb:2015) handle billion-scale graphs, and most are restricted to million-scale graphs due to memory and I/O constraints.
• The k-core model is the most widely used cohesiveness metric in CS, followed by k-truss and k-clique, each with distinct trade-offs in density and computational cost.
• There is a lack of publicly available code and datasets for most surveyed CS studies, hindering reproducibility and practical adoption.
• Distributed computation platforms (e.g., GraphX) and I/O-efficient disk-based algorithms are promising directions for scaling CS to real-world big graphs.
• No prior work addresses community search in signed graphs, multi-dimensional graphs, or uncertain graphs, indicating significant open research opportunities.