Novel Approach for Software Reliability Analysis Controlled with Multifunctional Machine Learning Approach

Abstract

Reliability engineering is distinguished from other fields by its focus on software. Models that forecast when things will go wrong are used to evaluate the reliability of a piece of software. Real-world issues might arise when it comes to reliability. Many computational algorithms have been developed in order to provide simple, trustworthy, and fast answers. Even though it doesn’t work, “Quality” is an important part of any programme that is often missing from software solutions. Using defect prediction models and object-oriented metrics, problem classes in software are found and used to measure its quality. Modeling software failure data and incorporating class-specific metrics peculiar to defective classes are used in this work to conduct an experimental investigation of defective classes. Machine learning and object-oriented metrics are used to do this. It is known how likely the classes of Marian Jureczko (MJ) Data sets are to go wrong. After looking at the Marian Jureczko Data set, it has been shown that RF gives the best accuracy and ROCAUC values. The RF model that was built is great.