Leveraging meta-heuristic algorithms for effective software fault prediction: a comprehensive study

Abstract In large-scale software development, the increasing complexity of software products poses a daunting challenge to maintaining software quality. Given this challenge, software fault prediction (SFP) is a critical endeavor for effective budgeting and refinement of the testing process. Quantitative insights into software quality gained through measurements are crucial in enabling accurate SFP. With the proliferation of software in various fields, ensuring software reliability throughout the software life cycle has become paramount. Anticipating software bugs, which have the potential to reduce software maintenance costs dramatically, is a key approach to improving software reliability. In this regard, using nature-inspired metaheuristic algorithms is promising because of their ability to predict future conditions and identify software anomalies. This study examines the potential of various meta-heuristic algorithms, particularly particle swarm optimization, genetic, ant colony optimization, cuckoo search, lion optimization, firefly, moth-flame, whale optimization, and artificial bee colony algorithms, in addressing the SFP challenge. The study outlines the challenging problems, compares approaches based on fundamental variables, and offers suggestions for future studies, providing a comprehensive and systematic analysis of these algorithms in the context of SFP.