GCN-Based Issues Classification in Software Repository

Abstract

Graph Convolutional Network (GCN) have demon- strated significant potential in various fields, particu- larly in classification tasks. This study introduces GCN- based methodology for classifying issues in software repositories, highlighting advancements in agile soft- ware development. Utilizing the dataset by Tawosi et al.(Tawosi et al. 2022), our research demonstrates the potential of GCNs to accurately categorize software is- sues into bugs, improvements, and tasks. Our results indicate a significant improvement in issue classifi- cation, especially for bugs. Additionally, we explore Fast Text GCN model, underlining their efficiency in handling dynamic, evolving datasets. This paper con- tributes to the fields of software engineering and ma- chine learning, offering novel insights into enhancing issue management in software projects.

Accessibility Summary:

In accordance with Title II regulations this content meets all points of exemption as Archived web content and/or Preexisting conventional electronic documents.