Implementation, trained models and result data for the paper Pairwise Multi-Class Document Classification for Semantic Relations between Wikipedia Articles (PDF on Arxiv). The supplemental material is available for download under GitHub Releases or Zenodo.

wget https://github.com/masukislam/DS8008_NLP/releases/download/Models/enwiki-20191101-pages-articles.weighted.10k.jsonl.bz2 bzip2 enwiki-20191101-pages-articles.weighted.10k.jsonl.bz2

wget https://github.com/masukislam/DS8008_NLP/releases/download/Model/train_testdata__4folds.tar.gz tar -xzf train_testdata__4folds.tar.gz

wget https://github.com/masukislam/DS8008_NLP/releases/download/Model/model_wiki.bert_base__joint__seq512.tar.gz tar -xzf model_wiki.bert_base__joint__seq512.tar.gz

## Demo

You can run a Jupyter notebook on Google Colab:

<a href="https://github.com/masukislam/DS8008_NLP/blob/master/demoRun.ipynb" target="_parent"><img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/></a>