Machine Learning Approaches in Contemporary Automatic Bug Triaging and Analysis of Research Gaps

The main aim of software development is to release the product with zero defects. Software development methodology starts from gathering the requirements till software testing maintenance. The testers perform a Software Testing process by comparing the actual and expected results. In the process of software testing whenever there is a mismatch between the expected and actual results, a bug is reported. Bug Triaging is a crucial process in software testing. Bug triaging refers to the process of getting the bug resolved at the right time by the right resource person The main idea of bug triaging is to get the bugs fixed within the timelines by the appropriate resource person. Bugs are addressed by the testing team using a bug report template. If a bug report is well-defined and easy to understand the developer finds it easier to understand and fix the bugs within the timelines. The bug report contains the bug’s characteristics including the description, ID, priority, steps to replicate the bug, hardware, and software configurations in which the bug was reported, and various other characteristics of the bug. Initially, the developers allocated the bugs using a manual distribution process. When there are hundreds or thousands of bugs reported in the software every month automatic bug triaging plays an efficient role in the timely resolution of the bugs. Bug triaging problems can be viewed as a machine learning application in which automatically the bugs are assigned to the appropriate developer. The main objective of Automatic Bug Triaging is to make a model that assigns bugs to appropriate developers with high accuracy to reduce the time of bug allocation to the developers. Various Machine Learning approaches have been used in the past years. In this paper, we will study various automatic bug-triaging approaches and discuss the findings and gaps of the discussed approaches. The main agenda of this research paper is to provide a detailed study related to Bug and Bug reports, analyze the various approaches to Automatic Bug Triaging and its practical implications, discuss the gaps in the existing approaches, and focus on the future scope and challenges related to automatic triaging. The customization, scalability, and adaptability of the machine learning techniques make it best suitable for Bug triaging purposes.