Sie befinden Sich nicht im Netzwerk der Universität Paderborn. Der Zugriff auf elektronische Ressourcen ist gegebenenfalls nur via VPN oder Shibboleth (DFN-AAI) möglich. mehr Informationen...
Software reliability prediction model based on PSO and SVM
Ist Teil von
2011 International Conference on Consumer Electronics, Communications and Networks (CECNet), 2011, p.5236-5239
Ort / Verlag
IEEE
Erscheinungsjahr
2011
Quelle
IEEE Electronic Library (IEL)
Beschreibungen/Notizen
Software reliability prediction classifies software modules as fault-prone modules and less fault-prone modules at the early age of software development. As to a difficult problem of choosing parameters for Support Vector Machine (SVM), this paper introduces Particle Swarm Optimization (PSO) to automatically optimize the parameters of SVM, and constructs a software reliability prediction model based on PSO and SVM. Finally, the paper introduces Principal Component Analysis (PCA) method to reduce the dimension of experimental data, and inputs these reduced data into software reliability prediction model to implement a simulation. The results show that the proposed prediction model surpasses the traditional SVM in prediction performance.