Bayesian network based software reliability prediction with an operational profile
Journal of Systems and Software
Hi-index | 0.00 |
Software defect curves describe the behavior of theestimate of the number of remaining software defects assoftware testing proceeds. They are of two possiblepatterns: single-trapezoidal-like curves or multiple-trapezoidal-like curves. In this paper we present some necessary and/or sufficient conditions for software defectcurves of the Goel-Okumoto NHPP model. Theseconditions can be used to predict the effect of thedetection and removal of a software defect on thevariations of the estimates of the number of remainingdefects. A field software reliability dataset is used tojustify the trapezoidal shape of software defect curves andour theoretical analyses. The results presented in thispaper may provide useful feedback information forassessing software testing progress and have potentials inthe emerging area of software cybernetics that exploresthe interplay between software and control.