Traceable link automatic recovery method based on multi-source information combination

The invention discloses a traceability link automatic recovery method based on multi-source information combination, which comprises the following steps: S1, selecting a target project, acquiring defect report data and code submission data, and constructing a target data set according to manually completed links; s2, for metadata information in the defect report, completing feature representation of the metadata information; for text information and code information in the defect report, obtaining vectorization embedding representation through a pre-training model; s3, splicing the obtained metadata feature vector x, text feature vector y and code feature vector z to obtain a multi-source fusion feature vector for link recovery; and S4, completing binary classification through a full connection layer by using the multi-source fusion feature vector, and determining whether a traceable link exists between the defect report and code submission. According to the method, the information in the link pair can be acc