Towards a constructive quality model: part 1: software quality modelling, measurement and prediction
Software Engineering Journal
A Pattern Recognition Approach for Software Engineering Data Analysis
IEEE Transactions on Software Engineering - Special issue on software measurement principles, techniques, and environments
Predictive Modeling Techniques of Software Quality from Software Measures
IEEE Transactions on Software Engineering - Special issue on software measurement principles, techniques, and environments
Rough Sets: Theoretical Aspects of Reasoning about Data
Rough Sets: Theoretical Aspects of Reasoning about Data
Line-crawling robot navigation: a rough neurocomputing approach
Autonomous robotic systems
An empirical study of using rule induction and rough sets to software cost estimation
Fundamenta Informaticae - Special issue on theory and applications of soft computing (TASC04)
Software defect prediction based on source code metrics time series
Transactions on rough sets XIII
| Hi-index | 0.00 |
This paper focuses on calibrating a rough neural network based on software complexity measurements and the corresponding number of changes required to bring a software product (either during development or during post-deployment) into compliance with project standards. A good predictive model for software maintenance that can estimate the number of changes that will allow the early identification of modules that are most likely to require extensive modifications. The results reported in this paper are limited to assessing prediction accuracy based on software engineering data obtained during product development. The Rough Set Exploration System (RSES) is used to derive training and testing sets that are used both by RSES and by a rough neural network toolset named MBnet to predict the number of software module changes needed to bring a module intro compliance with project standards. A comparison between MBnet and RSES in predicting the number of changes for a particular software module is also given.