Zero-shot Learning based Alternatives for Class Imbalanced Learning Problem in Enterprise Software Defect Analysis

Software defect reports are an important type of text data for enterprises as they provide actionable information for improving software quality. Identifying the software defect type automatically can greatly enhance and expedite defect management. Class imbalance is a real-life problem in enterprise software defect classification task and adversely affects the automation effort. We show that zero shot learning based technique can be a good alternative to the well-known supervised learning and SMOTE techniques.CCS CONCEPTS•Software and its engineering → Software defect analysis.

1 BACKGROUND AND MOTIVATION

Predicting the defect type of a software defect is valuable for enterprises as it supports proactive issue resolution, efficient bug triaging, optimization of testing resource allocation, and also helps in root cause analysis. Enterprises typically rely on supervised machine learning algorithms as well as their advanced variants proposed in literature (e.g., [6], [7]) to train models on historical data of software defect reports. These techniques get adversely impacted due to class imbalance challenge present in the real-life software defects. Class imbalance can be costly for an enterprise in terms of both financial resources and reputation. This impacts defect management in SDLC and inflates the maintenance cost.