Defect prediction aims at identifying software components that are likely to cause faults before a software is made available to the end-user. To date, this task has been modeled as a two-class classification problem, however its nature also allows it to be formulated as a one-class classification task. Previous studies show that One-Class Support Vector Machine (OCSVM) can outperform two-class classifiers for within-project defect prediction, however it is not effective when employed at a finer granularity (i.e., commit-level defect prediction). In this paper, we further investigate whether learning from one class only is sufficient to produce effective defect prediction model in two other different scenarios (i.e., granularity), namely cross-version and cross-project defect prediction models, as well as replicate the previous work at within-project granularity for completeness. Our empirical results confirm that OCSVM performance remain low at different granularity levels, that is, it is outperformed by the two-class Random Forest (RF) classifier for both cross-version and cross-project defect prediction. While, we cannot conclude that OCSVM is the best classifier, our results still show interesting findings. While OCSVM does not outperform RF, it still achieves performance superior to its two-class counterpart (i.e., SVM) as well as other two-class classifiers studied herein. We also observe that OCSVM is more suitable for both cross-version and cross-project defect prediction, rather than for within-project defect prediction, thus suggesting it performs better with heterogeneous data. We encourage further research on one-class classifiers for defect prediction as these techniques may serve as an alternative when data about defective modules is scarce or not available.

Replication studies in Software Engineering are indispensable for ensuring the reliability, generalizability, and transparency of research findings. They contribute to the cumulative growth of knowledge in the field and promote a scientific approach that benefits both researchers and practitioners. In this article, we report our experience replicating a recently published work proposing a Transformer-based approach for Agile Story Point Estimation” dubbed GPT2SP. GPT2SP was proposed with the intent of addressing the three limitations of a previous Deep Learning-based approach dubbed Deep-SE, and the results reported in the original study set GPT2SP as the new state-of-the-art. However, when we used the GPT2SP source code made publicly available by the authors of the original study, we found a bug in the computation of the evaluation measure and the reuse of erroneous results from previous work, which had unintentionally introduced biases in the GPT2SP's performance evaluation. In this study, we report on the results we obtained after fixing the issues present in the original study, which reveal that their results were in fact unintentionally inflated due to these issues and that despite advancements, challenges remain in providing accurate effort estimations for agile software projects.

In the last decade, several studies have explored automated techniques to estimate the effort of agile software development. We perform a close replication and extension of a seminal work proposing the use of Deep Learning for Agile Effort Estimation (namely Deep-SE), which has set the state-of-the-art since. Specifically, we replicate three of the original research questions aiming at investigating the effectiveness of Deep-SE for both within-project and cross-project effort estimation. We benchmark Deep-SE against three baselines (i.e., Random, Mean and Median effort estimators) and a previously proposed method to estimate agile software project development effort (dubbed TF/IDF-SVM), as done in the original study. To this end, we use the data from the original study and an additional dataset of 31,960 issues mined from TAWOS, as using more data allows us to strengthen the confidence in the results, and to further mitigate external validity threats. The results of our replication show that Deep-SE outperforms the Median baseline estimator and TF/IDF-SVM in only very few cases with statistical significance (8/42 and 9/32 cases, respectively), thus confounding previous findings on the efficacy of Deep-SE. The two additional RQs revealed that neither augmenting the training set nor pre-training Deep-SE play lead to an improvement of its accuracy and convergence speed. These results suggest that using semantic similarity is not enough to differentiate user stories with respect to their story points; thus, future work has yet to explore and find new techniques and features that obtain accurate agile software development estimates.

Background: Defect Prediction research aims at assisting software engineers in the early identification of software defect during the development process. A variety of automated approaches, ranging from traditional classification models to more sophisticated learning approaches, have been explored to this end. Among these, recent studies have proposed the use of ensemble prediction models (i.e., aggregation of multiple base classifiers) to build more robust defect prediction models. Aims: In this paper, we introduce a novel approach based on multi-objective evolutionary search to automatically generate defect prediction ensembles. Our proposal is not only novel with respect to the more general area of evolutionary generation of ensembles, but it also advances the state-of-the-art in the use of ensemble in defect prediction. Method: We assess the effectiveness of our approach, dubbed as Multi-objectiveEnsembleGeneration (MEG), by empirically benchmarking it with respect to the most related proposals we found in the literature on defect prediction ensembles and on multi-objective evolutionary ensembles (which, to the best of our knowledge, had never been previously applied to tackle defect prediction). Result: Our results show that MEG is able to generate ensembles which produce similar or more accurate predictions than those achieved by all the other approaches considered in 73% of the cases (with favourable large effect sizes in 80% of them). Conclusions: MEG is not only able to generate ensembles that yield more accurate defect predictions with respect to the benchmarks considered, but it also does it automatically, thus relieving the engineers from the burden of manual design and experimentation.

Background: Previous work has provided some initial evidence that Story Point (SP) estimated by human-experts may not accurately reflect the effort needed to realise Agile software projects. Aims: In this paper, we aim to shed further light on the relationship between SP and Agile software development effort to understand the extent to which human-estimated SP is a good indicator of user story development effort expressed in terms of time needed to realise it. Method: To this end, we carry out a thorough empirical study involving a total of 37,440 unique user stories from 37 different open-source projects publicly available in the TAWOS dataset. For these user stories, we investigate the correlation between the issue development time (or its approximation when the actual time is not available) and the SP estimated by human-expert by using three widely-used correlation statistics (i.e., Pearson, Kendall and Spearman). Furthermore, we investigate SP estimations made by the human-experts in order to assess the extent to which they are consistent in their estimations throughout the project, i.e., we assess whether the development time of the issues is proportionate to the SP assigned to them. Results: The average results across the three correlation measures reveal that the correlation between the human-expert estimated SP and the approximated development time is strong for only 7% of the projects investigated, and medium (58%) or low (35%) for the remaining ones. Similar results are obtained when the actual development time is considered. Our empirical study also reveals that the estimation made is often not consistent throughout the project and the human estimator tends to misestimate in 78% of the cases. Conclusions: Our empirical results suggest that SP might not be an accurate indicator of open-source Agile software development effort expressed in terms of development time. The impact of its use as an indicator of effort should be explored in future work, for example as a cost-driver in automated effort estimation models or as the prediction target.

Software defect prediction research has adopted various evaluation measures to assess the performance of prediction models. In this paper, we further stress on the importance of the choice of appropriate measures in order to correctly assess strengths and weaknesses of a given defect prediction model, especially given that most of the defect prediction tasks suffer from data imbalance.

Agile software development is nowadays a widely adopted practise in both open-source and industrial software projects. Agile teams typically heavily rely on issue management tools to document new issues and keep track of outstanding ones, in addition to storing their technical details, effort estimates, assignment to developers, and more. Previous work utilised the historical information stored in issue management systems for various purposes; however, when researchers make their empirical data public, it is usually relevant solely to the study's objective. In this paper, we present a more holistic and versatile dataset containing a wealth of information on more than half a million issues from 44 open-source Agile software, making it well-suited to several research avenues, and cross-analyses therein, including effort estimation, issue prioritization, issue assignment and many more. We make this data publicly available on GitHub to facilitate ease of use, maintenance, and extensibility.

In this paper, we introduce a novel approach to predictive modeling for software engineering, named Learning From Mistakes (LFM). The core idea underlying our proposal is to automatically learn from past estimation errors made by human experts, in order to predict the characteristics of their future misestimates, therefore resulting in improved future estimates. We show the feasibility of LFM by investigating whether it is possible to predict the type, severity and magnitude of errors made by human experts when estimating the development effort of software projects, and whether it is possible to use these predictions to enhance future estimations. To this end we conduct a thorough empirical study investigating 402 maintenance and new development industrial software projects. The results of our study reveal that the type, severity and magnitude of errors are all, indeed, predictable. Moreover, we find that by exploiting these predictions, we can obtain significantly better estimates than those provided by random guessing, human experts and traditional machine learners in 31 out of the 36 cases considered (86 percent), with large and very large effect sizes in the majority of these cases (81 percent). This empirical evidence opens the door to the development of techniques that use the power of machine learning, coupled with the observation that human errors are predictable, to support engineers in estimation tasks rather than replacing them with machine-provided estimates.