Heaviest and Densest Subgraph Computation for Binary Classification. A Case Study
摘要
This article presents a novel network-based data classification method. The classification problem is discussed as a graph theoretical problem. A real-valued data first is transformed to an undirected graph, and then the heaviest and densest subgraphs are detected based on an ant colony optimization approach. Numerical experiments conducted on a real-valued dataset show the potential of the proposed approach.