Discovering patterns that indicate software reliability provides valuable information to software project managers. Software Quality Classification (SQC) modeling is a methodology that can be used to discover reliability patterns of large software projects. However, the patterns found by SQC modeling may not be accurate and robust owing to insufficient information used in the training process. This study compares two genetic programming-based SQC models using different volumes of data. These data were extracted from seven different NASA software projects. The results demonstrate that combining data from different projects can produce more accurate and reliable patterns.
Liu, Yi; Adkins, Gerald; Yao, Jeng-Foung; and Williams, Gita
"Discovering Software Reliability Patterns Based On Multiple Software Projects,"
Journal of International Technology and Information Management: Vol. 16
, Article 6.
Available at: https://scholarworks.lib.csusb.edu/jitim/vol16/iss3/6