CODE: A Moving Window-Based Framework for Detecting Concept Drift in Software Defect Prediction

Concept Drift (CD) refers to the data distributions that may vary after a minimum stable period. CD negatively influences the models’ performance of software defect prediction (SDP) trained on past datasets when applied to the new datasets. Based on the previous studies of SDP, it is confirmed that the accuracy of prediction models negatively affect due to the changes of the data distributions. In this paper, a moving window-based concept drift detection (CODE) framework is proposed to detect CD in the chronological defect datasets and aimed to investigate the feasibility of alleviating CD in the chronological defect datasets. The proposed CODE framework consists of four steps, in which the first pre-processes the defect datasets and forms a cross-version (CV) chronological data, the second constructs the CV defect models, the third calculates the test statistics, and the fourth provides a hypothesis-test based CD detection method. In prior studies of SDP, it is observed that class rebalancing techniques modify the distribution of training datasets and improve the prediction performance of models. The ability of CODE framework is demonstrated by conducting experiments on 36 versions of 10 software projects. Some of the key findings are: (1) Up to 50% of the chronological defect datasets are drift-prone while applying the most popular classifiers used from the SDP literature. (2) The class rebalancing techniques had a positive impact on the prediction performance for CVDP by correctly classifying the CV defective modules and detected CD by up to 31% on the resampled datasets.