K.S. Maharasan * and V. Saravanan Pages 131 - 137 ( 7 )
Background: Data mining is an emerged and promising technology, and it utilized in the software engineering development process. It does not only enhance the accurateness, it also improves the reliability of the software. A software development application error is a fault or bug in a computer program. It produces wrong or unexpected outcomes. In traditional software development, faults manually triaged through a specialist developer, i.e., a human being triaged.
Methods: Manual fault triage takes long time and produce low accuracy for the huge amount of faults. To resolve above issues, An Efficient Integration of Instance and Aspect Preferment Algorithm (EIIAPA) is proposed to decrease the scale of fault report data concurrently and to enhance the accurateness of data.
Results & Conclusion: The proposed technique helps to validate & verify software application in effective way. Reduction of data on fault triage aims to construct a high-superiority set of fault data in the small-scale system through eliminating the fault report. To applying an algorithm, fault data set and attributes are extracted from every fault data set and train a predictive model based on the historical dataset. Based on Experimental evaluations, proposed methodology reduces 0.06 ET (Execution Time), and improves 0.5 P (Precision), 0.75 R (Recall), 0.39 F-M (F-Measure) and 5.07% (accuracy) compared than existing methodologies.
Software development process, data mining, An Efficient Integration of Instance and Aspect Preferment Algorithm (EIIAPA), compiler, fault triage, execution time, precision, recall, f-measure, accuracy.
Science and Humanities Faculty, Anna University, Chennai, Tamilnadu, Department of Computer Applications, Sri Venkateswara College of Computer Applications and Management, Coimbatore, Tamilnadu