2026-07-26T21:29:02Z https://meral.edu.mm/oai

oai:meral.edu.mm:recid/4584 2022-03-24T23:16:20Z 1582963302567:1597824175385 user-ucsy

Application of Neural Network for Predicting Software Development Faults using Object-Oriented Design Metrics Thwin, Mie Mie Thet Quah, Tong Seng In this paper, we present the application of neural networkfor predicting software development faults includingobject-oriented faults. Object-oriented metrics can be usedin quality estimation. In practice, quality estimation meanseither estimating reliability or maintainability. In thecontext of object-oriented metrics work, reliability istypically measured as the number of defects. Objectoriented design metrics are used as the independentvariables and the number of faults is used as dependentvariable in our study. Software metrics used include thoseconcerning inheritance measures, complexity measures,coupling measures and object memory allocationmeasures. We also test the goodness of fit of neuralnetwork model by comparing the prediction result forsoftware faults with multiple regression model. Our studyis conducted on three industrial real-tirne systems thatcontain a number of natural faults that has been reportedfor three years [1]. 2002-11-22 http://hdl.handle.net/20.500.12678/0000004584 https://meral.edu.mm/records/4584