Enhancing Software Reliability and Fault Detection Using Hybrid Brainstorm Optimization-Based LSTM Model

ABSTRACT

An essential attribute of software quality is software reliability. To achieve higher reliability, the testing phase with detected and corrected flaws is incorporated in the software development. The fault correction process (FCP) includes the fault detection process (FDP) to develop the software reliability growth model (SRGM). This is difficult to integrate because due to several reasons, including the effects of staffing levels and the interdependence of faults. It limits the applicability of the analytical model. Because of the adoption of data-driven methodologies such as Artificial Intelligence (AI) technology, no precise FCP and FDP assumptions are necessary. In this article, we proposed a hybrid long short-term memory (LSTM) with BrainStorm Optimization and Late Acceptance Hill Climbing (BSO-LAHC) algorithm of a stepwise prediction model for software fault detection and correction. The fault detection and correction procedure has great influence by considering the testing effort. While compared to the existing methods, the proposed hybrid with the BSO-LAHC algorithm demonstrated superior results by using Firefox and bug tracking system Bugzilla datasets. The proposed model’s effectiveness is confirmed via empirical study. Based on the Bugzilla and firefox datasets, the proposed mean square error performance is 1.92 and 21.44 respectively. Additionally, the proposed method is less expensive and takes less time to execute. In Bugzilla version 5.0.4, releases 2 and 3 had a determination coefficient of 99.2% and 98.9%, respectively. The FCP is 27% more effective than previous approaches, and the FDP is 32% more effective.

KEYWORDS:

• BrainStorm Optimization
• Late Acceptance Hill Climbing algorithm
• LSTM
• Software fault detection and correction
• Software reliability

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Lilly Raamesh

Lilly Raamesh is a professor of Information Technology at St Joseph’s College of Engineering since 2013. She has more than 20 years of teaching experience. She completed her PhD at Anna University, Chennai and her postgraduate studies (2000–2002) at Sathyabama Institute of Science and Technology and undergraduate (1990--1994) studies at Bharathidasan University. Her research interests lie in the area of software testing, data mining, machine learning and medical image processing. Dr Lilly Raamesh has attended roughly ten conferences and workshops and faculty development Programs. She has served as reviewer in the International Conference on Intelligent Computing, Smart Communication and Network Technologies-(ICICSCNT 2021).

S. Jothi

S Jothi (Soundaram Jothi) obtained her Bachelor of Engineering degree in computer science and engineering from Madurai Kamaraj University in 2003. Then she obtained her Master of Engineering in computer science and engineering from Annamalai University in 2005 and PhD in wireless sensor networks from Anna University Chennai in 2017. Currently, she is working as associate professor in the Department of Computer Science and Engineering at St Joseph’s College of Engineering, Chennai. Her specialization include wireless sensor networks, mobile ad-hoc networks, big data analysis and image processing. Her current research interests are machine learning and deep learning. She is a member in IEEE, life member in NCSSSASSOC, ISTE, CSI, ICSES, CRSI, IAENG, CSES, IACSIT and Fellow Member in ISRD. Email: [email protected]

S. Radhika

S Radhika (Radhika Sivashanmugam) is an associate professor, School of Electrical and Electronics Engineering of Sathyabama Institute of Science and Technology since 2006. She completed her PhD with research title “Design of adaptive filtering algorithms for acoustic echo cancellation application”. Her areas of research include Adaptive signal processing, system identification, echo cancellation and sparse signal processing. She has published several articles in international and national journals and conferences related to the adaptive filter algorithms. Email: [email protected]