opfdrop.blogg.se

We performed experiments on 10 open-source software systems from the PROMISE repository, which contain a total of 5,305.Problem prediction models can be used to direct test effort to defect-prone program code. To resolve this problem, cross-project defect prediction, which transfers a prediction model trained using data from one project to another, was proposed and is regarded as a new challenge in the area of defect prediction. The bug prediction dataset is a collection of models and metrics of software. In Proceedings of MSR 2010 (7th IEEE Working Conference on Mining Software Repositories), pp. (pf) then data resampling approaches should be avoided.An Extensive Comparison of Bug Prediction Approaches. However if the goal is to improve precision and reduce false alarm Resampling approaches for improved recall (pd) and g-measure prediction That software quality teams and researchers should consider applying data Significant positive effect of data resampling on CPDP performance, suggesting The authors' examined six defect prediction models on 34 datasets extractedįrom the PROMISE repository. Investigated and results are compared to approaches without data resampling.

Prediction performance of five oversampling approaches (MAHAKIL, SMOTE,īorderline-SMOTE, Random Oversampling, and ADASYN) and three undersamplingĪpproaches (Random Undersampling, Tomek Links, and Onesided selection) is Imbalance issue in CPDP, the authors assess the impact of data resamplingĪpproaches on CPDP models after the NN Filter is applied. Negative effects of class imbalance in the datasets. In the past, data resampling approaches have beenĪpplied to within-projects defect prediction models to help alleviate the A key challenge with defect-prediction datasets isĬlass imbalance, that is highly skewed datasets where non buggy modulesĭominate the buggy modules. Models using the Nearest Neighbour (NN) Filter approach have shown promising Projects are used to predict defects, has been proposed as a way to provideĭata for software projects that lack historical data. MacDonell, Jürgen Börstler Download PDF Abstract: Crossp-roject defect prediction (CPDP), where data from different software Authors: Kwabena Ebo Bennin, Amjed Tahir, Stephen G.