Abstract:
본 발명은, 안전한 유사 문서 검색 방법에 있어서, 제 1 문서 집합 및 제 2 문서 집합에 포함된 문서 벡터를 저차원으로 변환하는 단계, 상기 변환된 문서 벡터에 근거하여, 안전한 다자간 유사 문서 검색이 수행될 후보 집합을 획득하는 단계, 상기 후보 집합에 대한 안전한 다자간 유사 문서 검색을 수행하는 단계를 포함하는 안전한 유사 문서 검색 방법에 관한 것이다.
Abstract:
PURPOSE: A method for securely searching similar documents in two large document groups and an apparatus for the same are provided to reduce unnecessary calculation, thereby minimizing search time. CONSTITUTION: A communications unit (110) transmits and receives data with a first document group (10) and a second document group (20). A control unit (120) converts a document vector included in the first and second document groups into a low level, obtains a candidate group for performing a multilateral secure similar document search based on a converted document vector and performs the multilateral secure similar document search. The control unit obtains a first document frequency based on the first document group, obtains a second document frequency based on the second document group and selects data of a predetermined level based on the first document frequency and second document frequency. [Reference numerals] (10) First document group; (110,110-1,110-2) Transception unit; (120,120-1,120-2) Control unit; (130,130-1,130-2) Input unit; (20) Second document group