Ensemble Approach of ACOT and PSO for Predicting Software Reliability

Abstract

The importance on computer software has increased in recent decades. As computing systems become more numerous, complex, and deeply embedded in modern society, the need for systematic software development approaches tends to grow. System development problems that cause delays, increased costs, and/or failure to meet user needs are known as software crises. A systematic way to improve the quality of software by improving the development process can be incorporated into this challenging task. To predict software reliability, we proposed the Evolutionary Machine Learning algorithms ACOT, PSO, and a hybrid of ACOT and PSO. A comparison of our results with existing machine learning approaches such as neural networks and decision trees was also proposed. We used Root Mean Square Error and Normalized Root Mean Square Error to collect three software failure datasets to reinforce the demand besides software reliability.