A literature review on cross project defect prediction
- 1 October 2017
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 42 10, 680-685
Abstract
In the area of defect prediction most of the literature comprises of within project defect prediction. It is always not feasible to have the historical data of the similar projects for predictions. Therefore, CPDP (Cross Project Defect Prediction) as a subset of defect prediction in general has become a popular topic in research these days. In this paper, we present a systematic review of the CPDP. This paper summarizes the different methodologies used by various authors. A great amount of variation and heterogeneity is observed in the prediction process. The variations in the datasets, the learners, metric selection and the standard measure for comparisons are a challenge to determine the best practice for CPDP.Keywords
This publication has 23 references indexed in Scilit:
- Heterogeneous cross-company defect prediction by unified metric representation and CCA-based transfer learning2015
- Value-cognitive boosting with a support vector machine for cross-project defect predictionEmpirical Software Engineering, 2014
- Cross Company and within Company Fault Prediction using Object Oriented MetricsInternational Journal of Computer Applications, 2013
- Empirical evaluation of the effects of mixed project data on learning defect predictorsInformation and Software Technology, 2013
- Transfer learning for cross-company software defect predictionInformation and Software Technology, 2012
- Evaluating defect prediction approaches: a benchmark and an extensive comparisonEmpirical Software Engineering, 2011
- An investigation on the feasibility of cross-project defect predictionAutomated Software Engineering, 2011
- On the relative value of cross-company and within-company data for defect predictionEmpirical Software Engineering, 2009
- Predicting defect-prone software modules using support vector machinesJournal of Systems and Software, 2008
- A metrics suite for object oriented designIEEE Transactions on Software Engineering, 1994