GEval for KGRC-RDF-star

This repository is an extension of GEval.

This repository contains a (software) evaluation framework to perform evaluation and comparison on RDF-star graph embedding techniques.

How to use

If you want to run the tools, install Jupyter and run it on your host.

jupyter notebook

Example of Classification Task

run evaluation_rdfstar2vec.ipynb

Example of interpretation of the results

run tutorial_results_interpretation/results_interpretation_rdf-star2vec.py.ipynb

Tasks

The implemented tasks are:

• Machine Learning

  • Classification
  • Clustering

• Semantic tasks

  • Entity Relatedness
  • QT Similarity

Dataset

The gold standard datasets for evaluation were created from KGRC-RDF-star. Please see here.

Dependencies

The framework is tested to work with Python 3.8.3.

The required dependencies are: Numpy==1.14.0, Pandas==0.22.0, Scikit-learn==0.19.2, Scipy==1.1.0, H5py==2.8.0, unicodecsv==0.14.1.

License

The Apache license applies to the extended source code. The CC BY 4.0 applies to the gold standard datasets. The license of the original software can be found here. For source datasets, please check the license information here.