Information Retrieval Based Nearest Neighbor Classification for Fine-Grained Bug Severity Prediction

摘要: Bugs are prevalent in software systems. Some bugs critical and need to be fixed right away, whereas others minor their fixes could postponed until resources available. In this work, we propose a new approach leveraging information retrieval, particular BM25-based document similarity function, automatically predict the severity of bug reports. Our analyzes reports reported past along with assigned labels, recommends labels newly Duplicate utilized determine what report features, it textual, ordinal, or categorical, important. We focus on predicting fine-grained namely different Bugzilla including: blocker, critical, major, minor, trivial. Compared existing state-of-the-art study prediction, work by Menzies Marcus, our brings significant improvement.