Software　artifact　traceability　is　widely　recognized　as　an　essential　factor　for　effectively　managing　the　development　and　evolution　of　software　systems.　However,　such　traceability　links　are　usually　missed　in　practice　due　to　the　time　pressure.　Although　an　increasing　number　of　studies　have　been　carried　out　to　recover　such　links,　they　all　rely　on　calculating　the　textual　similarity　between　artifacts　without　appropriately　considering　the　context　of　each　artifact.　In　this　paper,　we　propose　a　novel　approach　to　recover　requirements　traceability　links　between　use　cases　and　code,　which　extends　Description-Embodied　Knowledge　Representation　Learning　(DKRL)　model　to　comprehensively　characterize　software　artifacts　by　embedding　both　text　information　and　interrelationships.　Such　meaningful　embeddings　are　then　used　to　train　traceability　link　classifiers　by　using　machine　learning　and　triple　classification　techniques.　Experimental　results　show　that　our　approach　is　superior　to　existing　approaches.　©　2020　Knowledge　Systems　Institute　Graduate　School.　All　rights　reserved.