Nature-inspired approaches in Software Fault Prediction

Nature-inspired approaches in Software Fault Prediction

Vidhi Jain1, Harshit Kumar2, Tushar Gupta3 , Nidhi Sengar4, Amita Goel5, Vasudha Bahl6

1,2,3,4,5,6 Department of Information Technology, Maharaja Agrasen Institute of Technology affiliated to Guru Gobind  Singh Indraprastha University, Rohini, Delhi

ABSTRACT

In software engineering, predicting software faults is a  crucial task for ensuring high software quality and reducing  costs. In recent years, nature inspired approaches have been  increasingly used in software fault prediction. In this paper,  we explore the effectiveness of six nature inspired  algorithms, namely Ant Colony, Particle Swarm  Optimization, Firefly, Bat, Harris Hawks, and Genetic  Algorithm, for software fault prediction. We evaluate the  algorithms using three commonly used datasets, JM1, CM1,  and PC1. Our experimental results show that nature inspired  approaches can effectively predict software faults, with some  algorithms performing better than others depending on the  dataset used. Our findings suggest that these approaches  have potential to be used as a practical and efficient means  for software fault prediction.

Keywords— nature inspired algorithms; PSO; Ant  Colony Optimization; Harris Hawks; Genetic Algorithm  (GA); python programming; Jupyter Notebook; confusion  matrix;