A Probabilistic Approach for Change Impact Prediction in Object-Oriented Systems.
Mustapha Kamel Abdi, Hakim Lounis, Houari A. Sahraoui · 2009
Abstract Several non probabilistic approaches were proposed in the literature to analyze and predict change impact in Object-Oriented (OO) systems. Different aspects were considered in these studies and several experiments were conducted to check some hypotheses. However, causality relation between software internal attributes and change impact still misses convincing explanations. In this paper, we propose a probabilistic approach using Bayesian networks to answer to this problematic of change impact analysis and prediction in OO systems. The built probabilistic model is tested on data extracted from a real system. The running of different scenarios on the network, globally confirm results already found in previous studies. 1