# Citation Prediction on arXiv citation network 169K papers (2022+)

Source: [Citation Prediction on arXiv citation network 169K papers (2022+)](<https://www.sota2.com/research/sota/citation-prediction-on-arxiv-citation-network-169k-papers-2022-plus>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-03-13T07:08:03.807Z

Evaluation of Citation Prediction performance on the arXiv citation network dataset comprising 169K papers published in 2022 and later, measured using ranking metrics like R@K and nDCG.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Citation Prediction |
| Dataset (including any stated split) | arXiv citation network 169K papers (2022+) |
| Metric | R@10 |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [R@10](<https://www.sota2.com/research/sota/citation-prediction-on-arxiv-citation-network-169k-papers-2022-plus.md?metric=R%4010>): Higher is better; unit: PERCENT.

- [R@20](<https://www.sota2.com/research/sota/citation-prediction-on-arxiv-citation-network-169k-papers-2022-plus.md?metric=R%4020>): Higher is better; unit: PERCENT.

- [nDCG@20](<https://www.sota2.com/research/sota/citation-prediction-on-arxiv-citation-network-169k-papers-2022-plus.md?metric=nDCG%4020>): Higher is better; unit: PERCENT.

- [MRR](<https://www.sota2.com/research/sota/citation-prediction-on-arxiv-citation-network-169k-papers-2022-plus.md?metric=MRR>): Higher is better; unit: PERCENT.

## Recorded results

| Method | R@10 | Result date | Report paper |
| --- | --- | --- | --- |
| GSS | 39.8 | Not recorded | 69a519582d603e6de3e3c6d9 |
| GAT+Euclidean | 34.1 | Not recorded | 69a519582d603e6de3e3c6d9 |
| BGE-Large | 33.8 | Not recorded | 69a519582d603e6de3e3c6d9 |
| Contriever | 32.5 | Not recorded | 69a519582d603e6de3e3c6d9 |
| SPECTER+FAISS | 31.2 | Not recorded | 69a519582d603e6de3e3c6d9 |
| Node2Vec | 28.7 | Not recorded | 69a519582d603e6de3e3c6d9 |

### GSS

R@10: 39.8. Result ID: 69b3b7bc47f34ce9249d3390.

[Reporting paper](<https://www.sota2.com/research/paper/69a519582d603e6de3e3c6d9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69a519582d603e6de3e3c6d9.md>)

Recorded evaluation configuration: {"Architecture":"GAT","Metric Factor Rank":"32","Layers":"3","Heads":"4","Embedding Dimension":"256"}. This record does not assert independent reproduction.

### GAT+Euclidean

R@10: 34.1. Result ID: 69b3b7bc47f34ce9249d338f.

[Reporting paper](<https://www.sota2.com/research/paper/69a519582d603e6de3e3c6d9.md>)

Recorded evaluation configuration: {"Architecture":"GAT","Distance Metric":"Fixed Euclidean distance (Li = I)","Layers":"3","Heads":"4","Embedding Dimension":"256"}. This record does not assert independent reproduction.

### BGE-Large

R@10: 33.8. Result ID: 69b3b7bc47f34ce9249d338e.

[Reporting paper](<https://www.sota2.com/research/paper/69a519582d603e6de3e3c6d9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74e31e209131cea64745.md>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### Contriever

R@10: 32.5. Result ID: 69b3b7bc47f34ce9249d338d.

[Reporting paper](<https://www.sota2.com/research/paper/69a519582d603e6de3e3c6d9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d4b1e209131cea38d50.md>)

[Implementation](<https://github.com/facebookresearch/contriever>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

### SPECTER+FAISS

R@10: 31.2. Result ID: 69b3b7bc47f34ce9249d338b.

[Reporting paper](<https://www.sota2.com/research/paper/69a519582d603e6de3e3c6d9.md>)

Recorded evaluation configuration: {"FAISS Index":"IndexFlatIP"}. This record does not assert independent reproduction.

### Node2Vec

R@10: 28.7. Result ID: 69b3b7bc47f34ce9249d338c.

[Reporting paper](<https://www.sota2.com/research/paper/69a519582d603e6de3e3c6d9.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa62a01e209131ce9fa578.md>)

[Implementation](<http://snap.stanford.edu/node2vec>)

Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
