CiteSeer Co-Citation

Download Dataset

Network Type

Simplicial ComplexHypergraph

Tags

domain: collaborationmetadata: node classes

Overview

The CiteSeer co-citation network is a hypergraph where nodes are scientific papers from the CiteSeer dataset and hyperedges are sets of papers that are co-cited together by another paper. Each node is labeled with the paper’s research field, chosen from six computer science topics: Agents, Artificial Intelligence (AI), Databases (DB), Information Retrieval (IR), Machine Learning (ML), and Human-Computer Interaction (HCI). This dataset is commonly used as a benchmark in hypergraph learning tasks.

Statistics

Interpretation

In the simplicial-complex interpretation, each co-citation set generates a simplex together with all of its subsets. The shape below is the full f-vector of that closure, with isolated papers retained as 0-simplices.

Nodes

3,312

Node Type

Scientific Paper

Node Label

Research Field

Active Vertices

1,458

Papers appearing in at least one co-citation set.

Maximal Simplices

832

Max Dimension

25

Node Label Distribution

6 unique labels · imbalance degree: 1.27

DimensionSimplex TypeCount
0Vertices3,312
1Edges4,307
2Triangles10,301
3Tetrahedra31,249
44-Simplices99,379
55-Simplices292,230
66-Simplices756,367
77-Simplices1,694,964
88-Simplices3,274,324
99-Simplices5,452,453
1010-Simplices7,835,423
1111-Simplices9,727,215
1212-Simplices10,436,420
1313-Simplices9,672,404
1414-Simplices7,730,853
1515-Simplices5,312,857
1616-Simplices3,124,739
1717-Simplices1,562,295
1818-Simplices657,801
1919-Simplices230,230
2020-Simplices65,780
2121-Simplices14,950
2222-Simplices2,600
2323-Simplices325
2424-Simplices26
2525-Simplices1
Total Simplices67,992,805