0 views collected around this technical thread.
This article introduces community detection as a fundamental graph algorithm, explains its basic concepts and types, compares it with clustering, discusses evaluation metrics like modularity, and reviews classic methods such as Louvain, node2vec‑based approaches, and the information‑theoretic Infomap algorithm.
The Louvain algorithm, a fast modularity‑maximizing community‑detection method, iteratively merges nodes into hierarchical super‑nodes, enabling anti‑fraud systems to uncover hidden collusive groups in weighted transaction graphs, thereby improving detection of fake orders, coupon abuse, and other illicit behaviors despite its iterative nature and streaming limitations.
This article explains how various graph algorithms—including connected components, label propagation, Louvain community detection, and Graph Convolutional Networks—are built on large‑scale user‑behavior graphs using Spark GraphX to detect and mitigate advertising fraud, detailing methodology, implementation, and experimental results.
The article explains the Louvain community‑detection algorithm, detailing its modularity‑maximizing objective, two‑step iterative process, efficient graph data structures, practical implementation considerations, and performance results on large‑scale graphs, providing a comprehensive guide for practitioners.
This article explains how graph‑based algorithms such as centrality measures, modularity optimization, Louvain, Infomap, node2vec and the comE framework can be applied to financial fraud detection and community discovery, detailing their principles, formulas, implementation steps and evaluation metrics.