This branch contains the implementations for the Heterogeneous Graph methods and pre-processing script. Please look at the other branches for the other improvement methods our team tried out.
This repository contains top solutions from WhoIsWho, the world’s largest manually-labeled name disambiguation benchmark.
WhoIsWho is the world’s largest manually-labeled name disambiguation benchmark, consisting of over 900,000 papers collected from AMiner database. The owners of each paper are annotated by skillful crowdworkers via the rigorous process.
We also propose two related representative name disambiguation tasks, Incremental Name Disambiguation and Name Disambiguation from Scratch, to comprehensively evaluate the capability of relevant name disambiguation methods.
Please go to from_scratch/rankx and incremental_name_disambiguation/rankx for reproduce details of Name Disambiguation from Scratch and Incremental Name Disambiguation, respectively.