JITBoost: Boosting Just-In-Time Defect Prediction Using Boolean Combination of Classifiers

We have meticulously prepared this comprehensive replication package to facilitate further investigation and validation of our groundbreaking study on Just-In-Time Software Defects Prediction (JIT-SDP) models. It serves as a valuable resource, offering intricate details about our research findings and in-depth analysis. The package is thoughtfully organized into separate, neatly compressed folders, ensuring easy accessibility and efficient data management. Notably, we have provided the ground truth for each dataset within the parsing results folder, offering a crucial reference point for researchers. Here is how the package is organized: Dataset: This section features distinct folders dedicated to each dataset utilized in our study. Notably, the data is meticulously categorized based on diverse features, encompassing process-related information, semantic details, and syntactic structures. Furthermore, for semantic information and syntactic structures, we have included CSV files containing the extracted features, accompanied by a binary file housing a comprehensive word dictionary. For instance, the project "Accumulo" encompasses two files: "accumulo_Deep_Data.csv," which encompasses the extracted features derived from commit messages and code changes, and "accumulo_dict.pickle," a valuable resource providing vocabulary banks for all project-related words. Results: This folder includes JIT-SDP results, including machine learning and deep learning models. It also contains the results of BCC algorithms (JIT-BBC, JIT-IBC, and JIT-WPIBC). Finally, you will find the results of all models used in this study. It is worth mentioning that our research findings have undergone rigorous scrutiny, with three independent researchers meticulously validating the analysis results to ensure unwavering accuracy and reliability. To use this replication package, follow these steps: Download the package from Zenodo and extract it to a location on your computer. Navigate to the folder corresponding to the JIT-SDP models you are interested in. Open the CSV files containing the JIT-SDP results for individual datasets and examine the data according to your research objectives