HiGitClass: Keyword-Driven Hierarchical Classification of GitHub Repositories

TL;DR

HiGitClass is a flexible, keyword-driven hierarchical classification framework for GitHub repositories that effectively integrates structured and unstructured data to improve topic-based search and analysis.

Contribution

The paper introduces HiGitClass, a novel framework that addresses challenges in automatic repository classification using minimal supervision and keyword hierarchies.

Findings

Outperforms existing weakly-supervised methods

Effectively integrates multi-modal signals

Handles supervision scarcity and bias

Abstract

GitHub has become an important platform for code sharing and scientific exchange. With the massive number of repositories available, there is a pressing need for topic-based search. Even though the topic label functionality has been introduced, the majority of GitHub repositories do not have any labels, impeding the utility of search and topic-based analysis. This work targets the automatic repository classification problem as keyword-driven hierarchical classification. Specifically, users only need to provide a label hierarchy with keywords to supply as supervision. This setting is flexible, adaptive to the users' needs, accounts for the different granularity of topic labels and requires minimal human effort. We identify three key challenges of this problem, namely (1) the presence of multi-modal signals; (2) supervision scarcity and bias; (3) supervision format mismatch. In recognition of these challenges, we propose the HiGitClass framework, comprising of three modules: heterogeneous information network embedding; keyword enrichment; topic modeling and pseudo document generation. Experimental results on two GitHub repository collections confirm that HiGitClass is superior to existing weakly-supervised and dataless hierarchical classification methods, especially in its ability to integrate both structured and unstructured data for repository classification.