earticle

영어

Software defects prediction was introduced to support development and maintenance activities such as improving the software quality through finding errors or patterns of errors early in the software development process. Software defects prediction is playing the role of maintenance facilitation in terms of effort, time and more importantly the cost prediction for software maintenance and evolution activities. In this research, software call graph model is used to evaluate its ability to predict quality related attributes in developed software products. As a case study, the call graph model is generated for several applications in order to represent and reflect the degree of their complexity, especially in terms of understandability, testability and maintenance efforts. This call graph model is then used to collect some software product attributes, and formulate several call graph based metrics. The extracted metrics are investigated in relation or correlation with bugs collected from customers-bug reports for the evaluated applications. Those software related bugs are compiled into dataset files to be used as an input to a data miner for classification, prediction and association analysis. Finally, the results of the analysis are evaluated in terms of finding the correlation between call graph based metrics and software products’ bugs. In this research, we assert that call graph based metrics are appropriate to be used to detect and predict software defects so the activities of maintenance and testing stages after the delivery become easier to estimate or assess.