Context. Software data collection precedes analysis which, in turn, requires data science related skills. Software defect prediction is hardly used in industrial projects as a quality assurance and cost reduction mean. Objectives. There are many studies and several tools which help in various data analysis tasks but there is still neither an open source tool nor standardized approach. Results. We developed Defect Prediction for software systems (DePress), which is an extensible software measurement, and data integration framework which can be used for prediction purposes (e.g. defect prediction, effort prediction) and software changes analysis (e.g. release notes, bug statistics, commits quality). DePress is based on the KNIME project and allows building workflows in a graphic, end-user friendly manner. Conclusions. We present main concepts, as well as the development state of the DePress framework. The results show that DePress can be used in Open Source, as well as in industrial project analysis.
Process metrics appear to be an effective addition to software defect prediction models usually built upon product metrics. We present a review of research studies that investigate process metrics in defect prediction. The following process metrics are discussed: Number of Revisions, Number of Distinct Committers, Number of Modified Lines, Is New and Number of Defects in Previous Revision. We not only introduce the definitions of the aforementioned process metrics but also present the most important results, recent advances and the summary regarding the use of these metrics in software defect prediction models, as well as the taxonomy of the analysed process metrics.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.