抽象的な

Advanced Matching Technique for Trustrace To Improve The Accuracy Of Requirement

S.Muthamizharasi, J.Selvakumar, M.Rajaram

During software evolution and maintenance, requirement traceability links become outdated because developers could not dedicate effort to updating them. So far, recovering these traceability links later is intimidated and costly task for developers. Earlier methods used information retrieval for recovering traceability links between free-text requirements and source code. However the accuracy is limited in such methods. To deal with this, the proposed system uses Relink, an automatic link recovery algorithm. The proposed approach is based on automatically learned feature criteria from explicit links. In addition the missing links can be found which affects defect prediction performance. We also assessed the impact of recovered links on software maintainability measurement, and established the result of ReLink affords considerably better accuracy than those of conventional heuristics.