Application of Principal Component Analysis in Software Quality Improvement
Statistical modeling technique has pivotal role in better understanding of the software development processes. Among them neural network techniques have enhanced predictive capability than most other statistical models. This paper explains the application of principal component analysis to neural network modeling as a way to improve predictability of neural network. The purpose of principal component analysis is to augment the performance of discriminant software quality models. The accurate neural training can be done by transferring the raw data into principal components. In this paper, the significance of principal components analysis is illustrated with the help of a commercial raw dataset and subsequently neural network modeling is described.