Statistical and regression analysis of impact of faults on object oriented software

Abstract

These days the Object-Oriented (OO) paradigm is used extensively in the development of software systems. The OO metrics can be employed to access the quality of these OO systems. Many metrics and metric suits have been developed by the scientists to access the quality of software. The properties of object-oriented designs can be measured using object-oriented metrics. With the help of metrics, estimations of project milestones can be attained more accurately. It is possible to develop software projects with nominal faults. A study shows that an estimated 42% on account of corrective maintenance can be saved by using object-oriented metrics. This paper assesses the capability of Object Oriented metrics to identify fault-proneness in Object Oriented software systems using statistical and regression analysis. Three projects from the NASA data set have been used to determine the applicability of object-oriented CK metrics. The CK metrics are used to predict the bugs in the class. The fault-proneness using OO metrics has been calculated. The information collected using this method will increase the quality and reliability of the OO software systems.