The　arrival　of　information　age　provides　a　platform　for　the　development　of　academic　field.　Knowledge　exchanges　between　disciplines　are　becoming　more　frequent,　boundaries　are　gradually　blurred,　interdisciplinary　has　become　an　inevitable　trend.　In　the　field　of　Scientometrics,　it　has　become　a　new　hotspot　to　study　the　interdisciplinarity　of　specific　documents　by　identifying　the　subject　characteristics　of　references.　However,　facing　a　massive　academic　documents,　how　to　identify　the　subject　characteristics　of　references　quickly　and　accurately　still　restricts　the　development　of　related　research.　In　this　paper,　we　use　the　reference　data　in　the　field　of　Gene　Editing　of　2015　from　Web　of　Science　core　database,　with　the　help　of　the　indicators　of　complex　network　link　prediction,　five　features　are　selected.　Base　on　this,　a　non-linear　Support　Vector　Machine　classification　model　is　constructed　to　judge　the　interdisciplinarity　of　citation.　Finally,　the　F1　value　obtained　by　5　fold　cross-validation　is　0.63,　which　indicators　that　the　indicators　can　distinguish　the　academic　citation　from　the　interdisciplinary　citation.　©　Springer　Nature　Switzerland　AG　2020.