loubnabnl / Link-Prediction-in-Citation-Graphs

Link prediction in a Citation Network using text and graph-based features. This work is part of the ALTEGRAD course in the MVA master.

ALTEGRAD Link prediction challenge:

Link prediction is the problem of predicting the existence of a link between two entities in a network. In this project, we worked on link prediction between scientific papers, the goal is to retrieve missing links in a citation network using text and graph-based features. This work was part of a challenge in the MVA ALTEGRAD course.

Setup

# Clone this repository
git clone https://github.com/loubnabnl/Link-Prediction-in-Citation-Graphs.git
cd Link-Prediction-in-Citation-Graphs/
# Install packages
pip install -r requirements.txt

Data and embeddings

You can download the data here. It contains information about the edges of the network, as well as the abstracts and authors of the papers. You can also download the embeddings we generated here. They must be respectively placed in the folders data/ and embeddings/.

Execution

The folder source/ contains scripts for generating node embeddings using Node2vec, abstract embeddings using two methods: SciBERT and Word2vec with a post-processing option to reduce the embedding size of SciBERT model. It also contains a script to generate author embeddings using NetMF on the nodes of an author graph we generate. The folder models/contain an MLP classification model, the other model: XGBoost, is direclty defined in main.py.

To execute the code, using our embeddings, you can run

python main.py