期刊名称：International Journal of Innovative Research in Science, Engineering and Technology
印刷版ISSN：2347-6710
电子版ISSN：2319-8753
出版年度：2014
期号：ICETS
页码：1307
出版社：S&S Publications
摘要：During software evolution and maintenance,requirement traceability links become outdatedbecause developers could not dedicate effort toupdating them. So far, recovering these traceabilitylinks later is intimidated and costly task fordevelopers. Earlier methods used informationretrieval for recovering traceability links betweenfree-text requirements and source code. However theaccuracy is limited in such methods. To deal with this,the proposed system uses Relink, an automatic linkrecovery algorithm. The proposed approach is basedon automatically learned feature criteria from explicitlinks. In addition the missing links can be found whichaffects defect prediction performance. We alsoassessed the impact of recovered links on softwaremaintainability measurement, and established theresult of ReLink affords considerably better accuracythan those of conventional heuristics.
