Context: In large-scale software development, a large number of bug reports are submitted during software development and maintenance. Practitioners need the ability to analyze this abundant data to make data-driven decisions about bug management tasks.

Objective: This thesis aims to utilize software analytics (SA) to support practitioners in bug management. The objectives of this thesis are (1) to identify and structure the knowledge on the use of SA for software engineering (SE) tasks and (2) to investigate and evaluate the practical application of SA to support practitioners in managing invalid bug reports (IBRs).

Method: We conducted a tertiary review and systematic mapping study to achieve the first objective and comparative experiments and two industrial case studies to achieve the second objective. Throughout the thesis work, we relied on a technology transfer model to guide the research and facilitate the adoption of ML techniques for the early identification of IBRs at the case company.

Results: We provide a comprehensive map of various SA applications for SE tasks and a decision matrix that can assist in selecting the most appropriate ML technique for bug report classification for a given context. Our results indicate that an ML technique can identify IBRs with acceptable accuracy at an early stage in practice. Furthermore, the results of an SA-based approach indicate that it can support practitioners in devising preventive measures for IBRs.

Conclusion: Through industrial validations, this thesis provides evidence of the usefulness of SA in bug management, particularly in supporting practitioners in managing IBRs in large-scale software development.