Title: Multi-label software bug categorisation based on fuzzy similarity

Authors: Rama Ranjan Panda; Naresh Kumar Nagwani

Addresses: Department of Computer Science and Engineering, National Institute of Technology – Raipur, Chhattisgarh, India ' Department of Computer Science and Engineering, National Institute of Technology – Raipur, Chhattisgarh, India

Abstract: The efficiency of the software depends on timely detection of bugs. For better quality and low-cost development bug fixing time should be minimised. Categorisation of software bugs helps to understand the root cause of software bugs and to improve triaging. As the software development approach is modular and multi-skilled, it is possible that one software bug can affect multiple modules, and multiple developers can fix newly reported bugs. Hence, a multi-label categorisation of software bugs is needed. Fuzzy similarity techniques can be helpful in understanding the belongingness of software bugs in multiple categories. In this paper a multi-label fuzzy similarity based categorisation technique is presented for effective categorisation of software bugs. Fuzzy similarity between a pair of bugs is computed and, based on a user defined threshold value, the bugs are categorised. Experiments are performed on software bug data sets, and the performance of the proposed classifier is evaluated.

Keywords: software bug mining; software bug classification; fuzzy similarity; multi-label classification; MLC; software bug repository.

DOI: 10.1504/IJCSE.2021.115645

International Journal of Computational Science and Engineering, 2021 Vol.24 No.3, pp.244 - 258

Received: 08 Jul 2020
Accepted: 11 Oct 2020

Published online: 04 Jun 2021 *

Full-text access for editors Access for subscribers Purchase this article Comment on this article