Selecting object-oriented source code metrics to improve predictive models using a parallel genetic algorithm

ABSTRACT

Predictive models can be used to discover potentially problematic components. Source code metrics can be used as input features to predictive models, however, there are many structural and design measures that capture related metrics of coupling, cohesion, inheritance, complexity and size. Feature selection is the process of identifying a subset of attributes that improves the performance of a predictive model. This paper presents a prototype that implements a parallel genetic algorithm as a search-based feature selection method that enhances a predictive model's ability to identify cognitively complex components in a Java application.

REFERENCES

• 1: Lionel C. Briand , John W. Daly , Jürgen Wüst, A Unified Framework for Cohesion Measurement in Object-OrientedSystems, Empirical Software Engineering, v.3 n.1, p.65-117, 1998 [doi>10.1023/A:1009783721306]
• 2: Lionel C. Briand , John W. Daly , Jürgen K. Wüst, A Unified Framework for Coupling Measurement in Object-Oriented Systems, IEEE Transactions on Software Engineering, v.25 n.1, p.91-121, January 1999 [doi>10.1109/32.748920]
• 3: Briand L.C. and Wuest J. Empirical Studies of Quality Models in Object-Oriented Systems. Advances in Computers, 56, pp. 97--166, 2002.
• 4: S. R. Chidamber , C. F. Kemerer, A Metrics Suite for Object Oriented Design, IEEE Transactions on Software Engineering, v.20 n.6, p.476-493, June 1994 [doi>10.1109/32.295895]
• 5: Khaled El-Emam, Object-oriented metrics: A review of theory and practice, Advances in software engineering, Springer-Verlag New York, Inc., New York, NY, 2002
• 6: David E. Goldberg, Genetic Algorithms in Search, Optimization and Machine Learning, Addison-Wesley Longman Publishing Co., Inc., Boston, MA, 1989
• 7: Repository of empirical software engineering data, URL: http://www.promisedata.org

INDEX TERMS

Primary Classification:
  D. Software
  D.2 SOFTWARE ENGINEERING
      D.2.8 Metrics
          Subjects: Product metrics

General Terms:
Design, Measurement, Reliability

Keywords:
genetic algorithm, object oriented metrics, predictive models