EvalNE: A framework for network embedding evaluation

In this paper we introduce EvalNE, a Python toolbox for network embedding evaluation. The main goal of EvalNE is to aid researchers and practitioners in performing consistent reproducible evaluations new methods, replicating existing evaluations, conducting benchmark studies. can evaluate models independently their programming language assess the quality learned representations through data visualization downstream tasks such as sign link prediction, reconstruction, node multi-label classification. streamlines evaluation by providing automation abstraction hyperparameter tuning model validation, edge sampling, node-pair computation, performance reporting. As command line tool, configuration files describe setup guarantee consistency reproducibility. an API, provides building blocks design any while minimizing risk errors.