Artifact selection using textual reports

摘要

Artifact selection by obtaining textual reports indicating potential defects in a program product. The textual reports are associated with artifacts of the program product. The textual reports and artifacts are automatically analyzed to estimate severity of potential errors in the artifacts of the program product, by utilizing a statistical model to estimate the severity based on content of the textual reports. The statistical model is trained with respect to a training dataset. A subset of the artifacts is determined based on the estimated severity.