Predicting Software Maintainability using Esemble Techniques and Stacked Generalization
octubre 30 @ 10:35 am - 11:05 am
The prediction of software maintainability has emerged as an important research topic to address industry expectations for reducing costs and, in particular, maintenance costs. In the last decades, many studies have used single techniques, statistical and machine learning, to predict software maintainability but no agreement has been reached on the technique that can achieve better prediction. Ensemble techniques, which combine two or more single techniques, have been in-vestigated in recent years to take advantage of single techniques. This study in-vestigates the use of ensemble techniques (homogeneous as well as heterogene-ous) on predicting maintainability in terms of change using stacked generalization method. To this end, well-known homogenous ensembles such as Bagging, Boosting, Extra Trees, Gradient Boosting, and Random Forest are investigated first. Then the stacked generalization combination rule is used to construct heter-ogeneous ensembles by combining the most accurate homogenous ensembles per dataset. The empirical results suggest that Gradient Boosting and Extra Trees are the best homogenous ensembles in all datasets, since they were ranked first and second in all datasets, respectively. Moreover, the findings of the evaluation of heterogeneous ensembles constructed using stacked generalization showed that they gave better prediction accuracy compared to all homogenous ensembles
Dr. Alain Abran is a professor of software engineering at the École de Technologie Supérieure (ETS) – Université du Québec, Montréal, Canada.
Dr. Abran has more than 20 years of industry experience in information systems development and software engineering, and 20 years of university teaching. He holds a PhD in electrical and computer engineering (1994) from École Polytechnique de Montréal (Canada) and Master’s degrees in Management Sciences (1974) and Electrical Engineering (1975) from the University of Ottawa (Canada). He is the chairman of the Common Software Measurement International Consortium (COSMIC) – www.cosmicon.com. He published Software Metrics and Software Metrology in 2010, Management of Software Maintenance1 in 2008, both at Wiley & IEEE CS Press, and co-edited the 2004 version of the Guide to the Software Engineering Body of Knowledge (www.swebok.org). His research interests include software productivity and estimation models, software quality, software measurement, functional size measurement methods, software risk management, and software maintenance management.
Sara ELMIDAOUI got a bachelor degree in mathematics and computer science in 2012 at Faculty of Sciences, University Ibn Zoher in Agadir, Morocco and a master degree in engineering design and application development from Faculty of Sciences and Technologies, University Hassan 1, Settat, Morocco in 2014. Sara is currently a Ph.D. student in Software Project Management team at National School of Computer Science & Systems Analysis (ENSIAS), Mohammed V University in Rabat, Morocco.