作者: Najeeb Ullah , Maurizio Morisio , Antonio Vetro
DOI: 10.1109/SEW.2012.26
关键词:
摘要: The purpose of this study is to compare the fitting (goodness fit) and prediction capability eight Software Reliability Growth Models (SRGM) using fifty different failure Datasets. These data sets contain defect collected from system test phase, operational phase (field defects) Open Source (OSS) projects. are modelled by SRGM (Musa Okumoto, Inflection S-Shaped, Goel Delayed Logistic, Gompertz, Yamada Exponential, Generalized Model). models chosen due their prevalence among many software reliability models. results can be summarized as follows o Fitting capability: Musa Okumoto fits all sets, but fit OSS datasets. Prediction S-Shaped best predictors for industrial Gompertz sets. performers on However happens only slightly more than 50%