Characterizing Software Test Case Behavior With Regression Models
B. Robbins Siemens PLM Software, Huntsville, AL, United States
Abstract
Testing modern software applications, such as those built on an event-driven paradigm, requires effective consideration of context. Model-based testing (MBT) approaches have been presented as an effective way to consider context, but effective MBT requires a high level of automation. Fully automated approaches too often lead to issues that threaten the very claims that MBT offers, such as generating test cases which are not fully executable and generating very large suites of test cases. In this chapter, I describe a new class of models for automated MBT workflows: predictive regression ...