Integrated probabilistic model for software quality prediction

The aim of the project is to develop a Bayesian net for software quality prediction. This model integrates multiple sources of data/konwledge. It adopts the most important advantages of existing software quality models, results of analyses of empirical data, and expert knowledge. The model provides new functionality by predicting various mutually related software quality factors using rigorous probability calculus. Developed model can serve as a basis for decision support system for software project managers. It offers a better picture of relationships between process and software quality factors. As a result project managers shall build better project plans make better decisions.

  • project number: N N111 291738
  • period – 5/2010-7/2012
  • role – manager and sole investigator
  • funding – Ministry of Science and Higher Education (Poland)

Links

Information on the predictive model

Results from empirical analysis

Questionnaire form that was used to calibrate the Bayesian network model with an expert knowledge

Related publications

[bibtex file=mypub.bib key=Radlinski_refsq2012,Radlinski_eknow2012,Radlinski_iccsa2011,Radlinski_aumcsi2011 format=luk]