The label quality of defect data sets has a direct influence on the reliability of defect prediction models. In this paper, we conduct a systematic study of inconsistent defect labels in multi-version-project defect data sets, i.e., many instances having the same source code but different labels over multiple versions of a software project. First, we report the phenomena of inconsistent labels by real examples and analyze their essence in the context of defect prediction. Then, we uncover the causes that lead to the occurrence of inconsistent labels for the representative label collection approaches. Finally, we investigate the actual influence of inconsistent labels on defect prediction models. We find that inconsistent labels in general exist in six multi-version-project defect data sets (either widely used or the most up-to-date in the literature) collected by diverse label collection approaches. In particular, inconsistent labels in a training data set significantly reduce the prediction performance of a model, while inconsistent labels in a test data set can lead to a considerable evaluation bias on the real performance. Therefore, we recommend that: on the one hand, researchers leverage our findings to make targeted methodological improvements on existing defect label collection approaches to reduce the generation of inconsistent labels; on the other hand, practitioners detect and exclude inconsistent labels in defect data sets to avoid their potential negative influence on defect prediction.

MOTIVATION: Scaling by sequencing depth is usually the first step of analysis of bulk or single-cell RNA-seq data, but estimating sequencing depth accurately can be difficult, especially for single-cell data, risking the validity of downstream analysis. It is thus of interest to eliminate the use of sequencing depth and analyze the original count data directly. RESULTS: We call an analysis method 'scale-invariant' (SI) if it gives the same result under different estimates of sequencing depth and hence can use the original count data without scaling. For the problem of classifying samples into pre-specified classes, such as normal versus cancerous, we develop a deep-neural-network based SI classifier named scale-invariant deep neural-network classifier (SINC). On nine bulk and single-cell datasets, the classification accuracy of SINC is better than or competitive to the best of eight other classifiers. SINC is easier to use and more reliable on data where proper sequencing depth is hard to determine. AVAILABILITY AND IMPLEMENTATION: This source code of SINC is available at https://www.nd.edu/∼jli9/SINC.zip. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

During the software system's maintenance and evolution, finding and removing software bugs is a very important part that consumes a large amount of money and effort.To analyze different bugs' character, it is very essential to know how long or which period of versions does the bug live in.In this study, we define version-based bug lifecycle and propose a text features based classification model to predict the versionlength of bug lifecycle.We collect 57000+ bugs from 10 well-know Apache Software Foundation projects to construct our dataset, and use the tf-idf method to collect our text features from bug report's summary and description.Our experimental results show that the text feature based method performs better than other baseline methods on 10 projects.The text feature based Naive Bayes classifiers outperforms all other methods with different features and classifiers.

Build prediction can reduce latency between continuous integration outcomes and the corresponding decisions, improving the efficiency of development team. Current build predictions are generally within-project, making it unavailable on those projects without enough build data. Cross-project prediction is the-state-of-art technique to solve the lack of training data on the studied projects by importing data from other projects. However, no previous study focuses on cross-project build predictions and checks the performance in the real world projects. This paper carries out an empirical study on the performance of cross-project build prediction with a wide range of 126 opensource projects under 6 common classifiers. In this paper, to select the training sets for cross-project prediction, we introduce two widely used data selection methods: Burak Filter based on build-level and Bellwether Strategy based on project-level. According to the results of our experiments, we have the following observations. Firstly, by the comparison between these two methods, we find that project-level selection (Bellwether strategy) performs better than build-level selection (Burak Filter). Furthermore, we observe that the prediction results can be improved by clustering the 126 studied projects into several smaller communities containing about 20-40 projects. And among 6 used classifiers, we find that decision tree classifier performs the best. Finally, by computing the optimal prediction results, we conclude that current selection methods still need to be improved to get close to the optimal prediction in cross-project build predictions.