Dual Principal Component Analysis in Software Reliability Studies

Abstract

Software usage has been occupying major parts in all the activities of individuals as well as organizations. Expectation of the end users about failure free operations of software is also increasing rapidly. Zhang and Pham defined 32 environmental factors for studying reliability of software and categorized them into five groups. Later Zhu et al. proposed to use information about three principal components extracted from 10 environmental factors only to study software reliability. It causes loss of information about the remaining 22 factors; two more environmental factors have been recommended as significant factors in a subsequent literature for studying reliability of software. This paper proposes a methodology to use the information about all the 34 factors through principal components, reducing the volume of information with relatively less amount of loss of information. Principal components are extracted in two-stage grouping of the 34 environmental factors. Information gained from the different stages of PCs is compared with Shannon information measure.

Keywords: Software reliability, environmental factor, clustering, principal components, Shannon entropy

Cite this Article

Loganathan A, Jeromia Muthuraj R. Dual Principal Component Analysis in Software Reliability Studies. Research & Reviews: Journal of Statistics. 2018; 7(1): 41s–46sp.