Fig. 3: The sequence shows the transformation of a dense network into a sparse graph through three distinct edge removal mechanisms.
From: Assessing the impact of sampling bias on node centralities in synthetic and biological networks
The node size indicates connectivity strength, demonstrating how different types of edge removal mechanisms affect network structure. In (a) Random Edge Removal (RER) simulates general missing information by randomly removing edges. In (b) Highly Connected Edge Removal (HCER) preferentially removes edges from high-degree nodes, representing overestimation of important node connections. In (c) Lowly Connected Edge Removal (LCER) focuses on removing edges from low-degree nodes, simulating overlooked peripheral connections”.
