Change-Proneness of Software Components
Ankita Urvashi, Anamika Chhabra · IOSR Journal of Computer Engineering · 2014
In this research paper, we have done a statistical study on the factors that influence the rate of change of piece of code (class or groups of class of working together).After conducting systematic literature survey with stake holders of a various software development.It was found that during the life cycle of the application development ,it undergoes many organic changes due to change in objectives of building the software, and it is now normal that the software remains in continuous change and measuring change in the software process is critical for the software to remain in use as well as for its relevant, hence ,In this paper we have found the trend on which the change-proneness can be modeled as function, a relationship equation that can help us to predict what will happen [good or bad] if a particular metric value changes.Previous work have assumed that all trends relationship are linear in nature, while dataset collected for shows it is highly nonlinear.Therefore, in our current research work we have found non-linear data fitting algorithm bi-square robust gives best possible results as it more accurate with realistic with trend and pass through most of test of significances ,thus we get a promising approach to measure change-proneness of the software development process .