Graph Comparison Gallery

This gallery compares each backbone method against a reference graph. Removed nodes are colored red and retained edges are drawn as thicker black lines.

For readability, isolates are removed from each backbone before plotting.

Les Miserables (Non-Bipartite Methods)

Glab Filter

• Module: hybrid

• Original graph: 77 nodes, 254 edges

• Backbone graph: 6 nodes, 5 edges

• Nodes removed: 71

Adamic Adar Index

• Module: proximity

• Original graph: 77 nodes, 254 edges

• Backbone graph: 26 nodes, 76 edges

• Nodes removed: 51

Cosine Backbone

• Module: proximity

• Original graph: 77 nodes, 254 edges

• Backbone graph: 27 nodes, 76 edges

• Nodes removed: 50

Dice Backbone

• Module: proximity

• Original graph: 77 nodes, 254 edges

• Backbone graph: 27 nodes, 76 edges

• Nodes removed: 50

Graph Distance Proximity

• Module: proximity

• Original graph: 77 nodes, 254 edges

• Backbone graph: 51 nodes, 76 edges

• Nodes removed: 26

Hub Depressed Index

• Module: proximity

• Original graph: 77 nodes, 254 edges

• Backbone graph: 29 nodes, 76 edges

• Nodes removed: 48

Hub Promoted Index

• Module: proximity

• Original graph: 77 nodes, 254 edges

• Backbone graph: 26 nodes, 76 edges

• Nodes removed: 51

Jaccard Backbone

• Module: proximity

• Original graph: 77 nodes, 254 edges

• Backbone graph: 27 nodes, 76 edges

• Nodes removed: 50

Lhn Local Index

• Module: proximity

• Original graph: 77 nodes, 254 edges

• Backbone graph: 39 nodes, 76 edges

• Nodes removed: 38

Local Path Index

• Module: proximity

• Original graph: 77 nodes, 254 edges

• Backbone graph: 24 nodes, 76 edges

• Nodes removed: 53

Neighborhood Overlap

• Module: proximity

• Original graph: 77 nodes, 254 edges

• Backbone graph: 22 nodes, 76 edges

• Nodes removed: 55

Preferential Attachment Score

• Module: proximity

• Original graph: 77 nodes, 254 edges

• Backbone graph: 33 nodes, 76 edges

• Nodes removed: 44

Resource Allocation Index

• Module: proximity

• Original graph: 77 nodes, 254 edges

• Backbone graph: 33 nodes, 76 edges

• Nodes removed: 44

Disparity Filter

• Module: statistical

• Original graph: 77 nodes, 254 edges

• Backbone graph: 10 nodes, 9 edges

• Nodes removed: 67

Ecm Filter

• Module: statistical

• Original graph: 77 nodes, 254 edges

• Backbone graph: 77 nodes, 254 edges

• Nodes removed: 0

• Validation warning: Filtered edge count equals the original graph. Re-test and validate this method's threshold or parameters.

Lans Filter

• Module: statistical

• Original graph: 77 nodes, 254 edges

• Backbone graph: 77 nodes, 109 edges

• Nodes removed: 0

Marginal Likelihood Filter

• Module: statistical

• Original graph: 77 nodes, 254 edges

• Backbone graph: 58 nodes, 70 edges

• Nodes removed: 19

Multiple Linkage Analysis

• Module: statistical

• Original graph: 77 nodes, 254 edges

• Backbone graph: 77 nodes, 109 edges

• Nodes removed: 0

Noise Corrected Filter

• Module: statistical

• Original graph: 77 nodes, 254 edges

• Backbone graph: 70 nodes, 98 edges

• Nodes removed: 7

Doubly Stochastic Filter

• Module: structural

• Original graph: 77 nodes, 254 edges

• Backbone graph: 77 nodes, 109 edges

• Nodes removed: 0

Edge Betweenness Filter

• Module: structural

• Original graph: 77 nodes, 254 edges

• Backbone graph: 77 nodes, 77 edges

• Nodes removed: 0

Global Sparsification

• Module: structural

• Original graph: 77 nodes, 254 edges

• Backbone graph: 44 nodes, 102 edges

• Nodes removed: 33

Global Threshold Filter

• Module: structural

• Original graph: 77 nodes, 254 edges

• Backbone graph: 58 nodes, 157 edges

• Nodes removed: 19

H Backbone

• Module: structural

• Original graph: 77 nodes, 254 edges

• Backbone graph: 19 nodes, 22 edges

• Nodes removed: 58

High Salience Skeleton

• Module: structural

• Original graph: 77 nodes, 254 edges

• Backbone graph: 77 nodes, 76 edges

• Nodes removed: 0

Maximum Spanning Tree Backbone

• Module: structural

• Original graph: 77 nodes, 254 edges

• Backbone graph: 77 nodes, 76 edges

• Nodes removed: 0

Metric Backbone

• Module: structural

• Original graph: 77 nodes, 254 edges

• Backbone graph: 77 nodes, 163 edges

• Nodes removed: 0

Modularity Backbone

• Module: structural

• Original graph: 77 nodes, 254 edges

• Backbone graph: 44 nodes, 72 edges

• Nodes removed: 33

Node Degree Filter

• Module: structural

• Original graph: 77 nodes, 254 edges

• Backbone graph: 60 nodes, 237 edges

• Nodes removed: 17

Planar Maximally Filtered Graph

• Module: structural

• Original graph: 77 nodes, 254 edges

• Backbone graph: 77 nodes, 162 edges

• Nodes removed: 0

Primary Linkage Analysis

• Module: structural

• Original graph: 77 nodes, 254 edges

• Backbone graph: 77 nodes, 69 edges

• Nodes removed: 0

Strongest N Ties

• Module: structural

• Original graph: 77 nodes, 254 edges

• Backbone graph: 77 nodes, 113 edges

• Nodes removed: 0

Ultrametric Backbone

• Module: structural

• Original graph: 77 nodes, 254 edges

• Backbone graph: 77 nodes, 118 edges

• Nodes removed: 0

Local Degree

• Module: unweighted

• Original graph: 77 nodes, 254 edges

• Backbone graph: 77 nodes, 135 edges

• Nodes removed: 0

Lspar

• Module: unweighted

• Original graph: 77 nodes, 254 edges

• Backbone graph: 77 nodes, 136 edges

• Nodes removed: 0

Sparsify

• Module: unweighted

• Original graph: 77 nodes, 254 edges

• Backbone graph: 77 nodes, 136 edges

• Nodes removed: 0

Davis Southern Women (Bipartite Methods)

Fdsm

• Module: bipartite

• Original graph: 18 nodes, 153 edges

• Backbone graph: 11 nodes, 9 edges

• Nodes removed: 7

Fixedcol

• Module: bipartite

• Original graph: 18 nodes, 153 edges

• Backbone graph: 7 nodes, 9 edges

• Nodes removed: 11

Fixedfill

• Module: bipartite

• Original graph: 18 nodes, 153 edges

• Backbone graph: 7 nodes, 9 edges

• Nodes removed: 11

Fixedrow

• Module: bipartite

• Original graph: 18 nodes, 153 edges

• Backbone graph: 16 nodes, 17 edges

• Nodes removed: 2

Sdsm

• Module: bipartite

• Original graph: 18 nodes, 153 edges

• Backbone graph: 6 nodes, 3 edges

• Nodes removed: 12