Fault Prediction Modeling using Object-Oriented Metrics: An Empirical Study
Pages : 254-259
Software testing is an area where software products are examined through a series of verification and validation processes respectively. This phase of software development carries out the process of detection and removal of software faults. But this detection and removal of faults together consume up to 60% of project budget (Beizer, 1990). Applying equal testing and verification efforts to all parts of a software system becomes cost-prohibited. Software Fault Proneness is a key factor for monitoring and controlling the quality of software. By comparing the distribution of faults (Fault Proneness) and the amount of faults found with testing (Software faultiness), the effectiveness of analysis and testing can be judged easily. Detecting a fault prone code early, within software life-cycle phase, allows for the code to be fixed at minimum costs; thus a good fault prediction helps to lower down the development and maintenance costs. In this project, software quality estimation, using various classifiers is performed where some metrics are the inputs and its quality attributes bring the output. Empirical study of these classifiers then judges the quality of the software being developed.
Keywords: Software Testing, Classification Algorithms, bugs, Weka 3.6.9 tools, Performance analysis.
Article published in International Journal of Current Engineering and Technology, Vol.4,No.1 (Feb- 2014)