/GeneticFlow

GeneticFlow: Graph Signatures for Scholar Impact Evaluation using Self-Citations

Primary LanguagePythonMIT LicenseMIT

Folder

[Advisor-advisee_Relationship_Detection] :

We provide the data and code, which implements our algorithm for advisor-advisee detection task.

[Alphabetically_Ordered_Paper_Detection] :

We provide a python file, which implements our method for alphabetically ordered paper detection.

[Inference_of_Extend-type_Citations] :

Within this folder, extra-trees classifier are trained with labeled data, for the detection task of extend-type citations.

[Award_Inference_Task] :

Based on award author information, we conduct award inference experiments, using GNN model, which embed the GeneticFlow graph (or other alternative graphs) and node/edge relevancy into vector representation for downstream academic impact evaluation tasks.