Abstract:
Defect tracking using computational intelligence
methods is used to predict software readiness in this study. By
comparing predicted number of faults and number of faults
discovered in testing, software managers can decide whether the
software are ready to be released or not.
Our predictive models can predict: (i) the number of faults
(defects), (ii) the amount of code changes required to correct a
fault and (iii) the amount of time (in minutes) to make the changes
in respective object classes using software metrics as independent
variables. The use of neural network model with a genetic training
strategy is introduced to improve prediction results for estimating
software readiness in this study.
Our prediction model is divided into three parts: (1) prediction
model for Presentation Logic Tier software components (2)
prediction model for Business Tier software components and (3)
prediction model for Data Access Tier software components.
Existing object-oriented metrics and complexity software metrics
are used in the Business Tier neural network based prediction
model. New sets of metrics have been defined for the Presentation
Logic Tier and Data Access Tier. These metrics are validated
using two sets of real world application data, one set was collected
from a warehouse management system and another set was
collected from a corporate information system.
