Bayesian Belief Networks (BBN)

Bayesian Belief Networks (BBN) is a hybrid estimation method. It represents a model-based, parametric estimation method that implements a define-your-own-model approach. Actually, for the purpose of software effort estimation, the method adapts the concept of Bayesian Networks, which has been evolving for many years in probability theory. The approach was recently adapted to software estimation due to its ability to combine knowledge based on quantitative measurement data and human judgment into intuitive graphical models with a sound theoretical basis. These applications include estimation of software development productivity (Stewart 2002; Pendharkar et al. 2005) and effort (Moses and Clifford 2000; Stamelos et al. 2003a; Mendes 2007; Hamdan et al. 2009). This ability is particularly attractive in the software engineering context where measurement data are scarce and much knowledge is hidden in the heads of human experts.

[1]  Emilia Mendes,et al.  Bayesian Network Models for Web Effort Prediction: A Comparative Study , 2008, IEEE Transactions on Software Engineering.

[2]  Parag C. Pendharkar,et al.  A probabilistic model for predicting software development effort , 2003, IEEE Transactions on Software Engineering.

[3]  Ioannis Stamelos,et al.  On the use of Bayesian belief networks for the prediction of software productivity , 2003, Inf. Softw. Technol..

[4]  John Moses.  Learning how to improve effort estimation in small software development companies , 2000, Proceedings 24th Annual International Computer Software and Applications Conference. COMPSAC2000.

[5]  B. Stewart.  Predicting project delivery rates using the Naive-Bayes classifier , 2002, J. Softw. Maintenance Res. Pract..

[6]  Emilia Mendes.  A Comparison of Techniques for Web Effort Estimation , 2007, ESEM 2007.

[7]  Noura Abbas.  Agile Software Assurance: An Empirical Study , 2007, ESEM 2007.

[8]  K. Hamdan,et al.  A bayesian belief network cost estimation model that incorporates cultural and project leadership factors , 2009, 2009 IEEE Symposium on Industrial Electronics & Applications.