Two definitions of software defect prediction

1. Software defect prediction is to predict the possible defects of the software as soon as possible based on the metric information extracted from the software product and the defects that have been discovered. Based on the prediction results, testing and verification resources can be reasonably allocated.

2. Software defect prediction is an important direction of empirical software engineering. Through the mining and analysis of historical data accumulated in the process of software development, a measurement element with strong correlation with software defects is designed, and program modules are measured, and then statistics are used. Or machine learning technology to build a defect prediction model to achieve the prediction of the defect tendency, defect severity, or defect number distribution of the target software module.