CoreScope: Graph Mining Using k-Core Analysis - Patterns, Anomalies, and Algorithms

Overview

In this work, we presents three empirical patterns related to k-cores in real-world graphs and their applications to anomaly detection, streaming algorithm design, and influential spreaders identification.

Papers

• CoreScope: Graph Mining Using k-Core Analysis - Patterns, Anomalies and Algorithms
  Kijung Shin, Tina Eliassi-Rad, and Christos Faloutsos
  IEEE International Conference on Data Mining (ICDM) 2016, Barcelona, Spain
  [PDF] [Supplementary Document] [BIBTEX]
  
  
• Patterns and Anomalies in k-Cores of Real-World Graphs with Applications
  Kijung Shin, Tina Eliassi-Rad, and Christos Faloutsos
  Knowledge and Information Systems (KAIS), vol. 54, no. 3, pp. 677-710, March 2018
  [PDF] [BIBTEX]
  
  

Code

CoreScope v2.0 [Github Repository] includes

• Core-A: an anomaly detection algorithm based on coreness
• Truss-A: an anomaly detection algorithm based on trussness
• Core-D: a streaming algorithm for degeneracy
• Core-S: an influential spreader detection method based on the structure of degeneracy-cores

Datasets

People

• Kijung Shin (Carnegie Mellon University)
• Tina Eliassi-Rad (Northeastern University)
• Christos Faloutsos (Carnegie Mellon University)