Ensemble learning methods have been used to enhance the reliability of defect prediction models. However, there is an inconclusive stability of a single method attaining the highest accuracy among various software projects. This work aims to improve the performance of ensemble-learning defect prediction among such projects by helping select the highest accuracy ensemble methods. We employ bandit algorithms (BA), an online optimization method, to select the highest-accuracy ensemble method. Each software module is tested sequentially, and bandit algorithms utilize the test outcomes of the modules to evaluate the performance of the ensemble learning methods. The test strategy followed might impact the testing effort and prediction accuracy when applying online optimization. Hence, we analyzed the test order's influence on BA's performance. In our experiment, we used six popular defect prediction datasets, four ensemble learning methods such as bagging, and three test strategies such as testing positive-prediction modules first (PF). Our results show that when BA is applied with PF, the prediction accuracy improved on average, and the number of found defects increased by 7% on a minimum of five out of six datasets (although with a slight increase in the testing effort by about 4% from ordinal ensemble learning). Hence, BA with PF strategy is the most effective to attain the highest prediction accuracy using ensemble methods on various projects.

To enhance the reliability of software, it is important is to review all software artifacts (e.g., design documents) to remove defects as earlier as possible. There are various review methods available, and project managers face the challenge of choosing a suitable method for their current projects. One of approaches to support the selection of review methods is to evaluate review methods beforehand, to identify the most effective method on average. However, past studies have not evaluated review methods thoroughly as the process can be time-consuming. We propose a bandit-algorithm (BA) based method to evaluate and then dynamically select a suitable review method (from a list of candidates). In our experiments, we assume that the proposed method is applied to design document review on basic design phase. We performed experiments based on a simulation, instead of using an actual dataset. On our simulation, when a review method is selected by our BA method, productivity (i.e., total development time) was improved by about 1.25 times, and it was the second highest among candidates of review methods.

It is important for software development companies to consider the job satisfaction and turnover intention of employees. To simply explain the factors related to them, this study focused on future perspective index (FPI). The FPI is assumed to relate positively to satisfaction and negatively to turnover. A preliminary analysis improved FPI and allowed for better explained satisfaction and intention than native FPI.

Background: Selecting a suitable feature reduction technique. when building a defect prediction model, can be challenging. Different techniques can result in the selection of different independent variables which have an impact on the overall performance of the prediction model. To help in the selection, previous studies have assessed the impact of each feature reduction technique using different datasets. However, there are many reduction techniques, and therefore some of the well-known techniques have not been assessed by those studies. Aim: The goal of the study is to select a high-accuracy reduction technique from several candidates without preliminary assessments. Method: We utilized bandit algorithm (BA) to help with the selection of best features reduction technique for a list of candidates. To select the best feature reduction technique, BA evaluates the prediction accuracy of the candidates, comparing testing results of different modules with their prediction results. By substituting the reduction technique for the prediction method, BA can then be used to select the best reduction technique. In the experiment, we evaluated the performance of BA to select suitable reduction technique. We performed cross version defect prediction using 14 datasets. As feature reduction techniques, we used two assessed and two non-assessed techniques. Results: Using BA, the prediction accuracy was higher or equivalent than existing approaches on average, compared with techniques selected based on an assessment. Conclusions: BA can have larger impact on improving prediction models by helping not only on selecting suitable models, but also in selecting suitable feature reduction techniques.

Background: defect prediction model is built using historical data from previous versions/releases of the same project. However, such historical data may not exist in case of newly developed projects. Alternatively, one can train a model using data obtained from external projects. This approach is known as cross-project defect prediction (CPDP). In CPDP, it is still difficult to utilize external projects' data or decide which particular project to use to train a model. Aim: to address this issue, we apply bandit algorithm (BA) to CPDP in order to select the most suitable training project from a set of projects. Method: BA-based prediction iteratively reselects the project after each module is tested, considering the accuracy of the predictions. As baselines, we used simple CPDP methods such as training a model with randomly selected project. All models were built using logistic regression. Results: We experimented our approach on two datasets (NASA and DAMB, with a total of 12 projects). The BA-based defect prediction models resulted in, on average, a higher accuracy (AUC and F1 score) than the baselines. Conclusion: in this preliminarily study, we demonstrate the feasibility of using BA in the context of CPDP. Our initial assessment shows that the use BA for predicting defects in CPDP is promising and may outperform existing approaches.

Background: Machine learning models are sometimes embedded in software to implement the required functions. As a result, non-experts in machine learning are becoming familiar with the models. However, the interpretability of the built models is often low in machine learning, such as deep learning, and the recognition process of such models is very different from that of humans. Therefore, it is not easy for novice users, such as end-users and beginners, to anticipate the behavior of models that they will use or build. Aim: We assist novice users to realize an aspect of the behavior of machine learning models relating to robustness intuitively. Method: We formalized and evaluated quiz-based analysis, which is often applied by practitioners to test the robustness of machine learning models arbitrarily. To generate test cases of the models, the analysis converts images towards the boundary of classification for both machine learning and humans. It can be regarded as a type of boundary value analysis of software development. Results: In the experiment, we evaluated whether the analysis quantitatively clarified the aspects of the models. The analysis clarified the robustness of the model for image conversion and misclassification quantitatively. Conclusion: The analysis is expected to enlighten novice users on the behavior of machine learning models. This may promote behavioral changes in the evaluation of models for novice users.

Abstract Although many software development projects have moved their developer discussion forums to generic platforms such as Stack Overflow, Eclipse has been steadfast in hosting their self-supported community forums. While recent studies show forums share similarities to generic communication channels, it is unknown how project-specific forums are utilized. In this paper, we analyze 832,058 forum threads and their linkages to four systems with 2,170 connected contributors to understand the participation, content and sentiment. Results show that Seniors are the most active participants to respond bug and non-bug-related threads in the forums (i.e., 66.1% and 45.5%), and sentiment among developers are inconsistent while knowledge sharing within Eclipse. We recommend the users to identify appropriate topics and ask in a positive procedural way when joining forums. For developers, preparing project-specific forums could be an option to bridge the communication between members. Irrespective of the popularity of Stack Overflow, we argue the benefits of using project-specific forum initiatives, such as GitHub Discussions, are needed to cultivate a community and its ecosystem.

We consider Free/libre and open source software (FLOSS) as a common pool resource (CPR). In economics, CPRs are frequently associated with markets, and it is reported that without appropriate agreement, monitoring and sanction, the resource will be overused. Toward building sustainable communities in FLOSS development, we first study our car-sharing experiment at NAIST, as a common pool resource management. We report the details of the car uses in our experiment, and describe the design of our new system to make better CPR management.