Premium
Software reliability model selection
Author(s) -
Khoshgoftaar Taghi M.,
Woodcock Timothy G.
Publication year - 1992
Publication title -
quality and reliability engineering international
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.913
H-Index - 62
eISSN - 1099-1638
pISSN - 0748-8017
DOI - 10.1002/qre.4680080509
Subject(s) - akaike information criterion , reliability engineering , software quality , reliability (semiconductor) , selection (genetic algorithm) , computer science , software reliability testing , model selection , software , data mining , software development , machine learning , engineering , physics , quantum mechanics , programming language , power (physics)
No one software reliability growth model has ever been shown to work well in all circumstances. This paper presents our evaluation results for two case studies in which the Akaike information criterion (AIC) was used. The AIC not only selects the best model among several reliability models, but also possess favourable properties that practitioners like to see in their software reliability modelling practices. These properties include simplicity, accuracy and ease of application. We propose using the Akaike information criterion to select the best model for each software system.