Database Reference
In-Depth Information
5900 PRs related to
sw-bug
. Any potential rule associated to
change-request
can be
heavily affected due to the presence of large size group with other values. Again the use
of additional data sources together with the PR data set can rectify this problem.
We also attempted to use neural networks, there was no significant improvement in
accuracy. In addition, it was difficult to interpret the outputs without the use of rule
extraction techniques.
4 Conclusion
This paper attempts to show the use of data mining as one of the efficient methods to
improve the SDLC process. Data mining can assist software developers by
automating some of the development tasks. For example, the fundamental idea behind
object-oriented programming is the re-use of components and linkage of objects
through class instantiation, polymorphism and abstraction calls. Data mining can help
to realise this concept more efficiently.
This paper also explored the use of data mining on a set of data collected during
the SDLC process under a real software business environment. Some useful rules are
inferred on the time consumption to fix a Problem Report and the relationship
between the content and the type of the PR. These rules are in the form of
associations, decision trees and semantic trees. The result may help developers in
problem reasoning, and project leaders in estimation and planning.
Results of this application indicate that DM has capacity to improve the quality and
efficiency of the software development process, even though the scale of this DM task
was limited. It will be interesting to apply data mining to different phases of software
development such as software quality data and integration of various data and
knowledge at multiple levels.
As in many other domains, the benefits and capabilities brought by data mining in
SDLC are worth of further investigations.
Acknowledgment
We will sincerely like to thank Mihir Shah, Parita Choksi and Magnus Haugaasen
(ITB239: Enterprise Data Mining students at QUT) for conducting a short literature
review on DM in SE domain. We will also like to show our gratitude to Dr Anurag
Nayak a Senior IT Consultant for providing answers to SDLC related questions.
References
1.
CBA, http://www.comp.nus.edu.sg/~dm2/
2.
C5.0, http://www.rulequest.com/see5-info.html
3.
TextMiner, http://www.megaputer.com/company/index.html
4.
EMERALD, http://www.graphicsillustrated.com/reliametrics/products/tools.html
5.
Alvarez-Macias J., J. Mata-Vazquez and J. Riquelme-Santos,
Data Mining for the
Management of Software Development Process,
IJSEKE, Vol. 14, Issue 6 (2004) 665-695.
Search WWH ::
Custom Search