Papers with Code 아카이브

Blocking

태스크의 하위 분류와 논문·코드·평가·기법·데이터셋을 탐색하세요.

하위 태스크
0
평가표
5
보존된 평가 기록
18

Entity resolution (also known as entity matching, record linkage, or duplicate detection) is the task of finding records that refer to the same real-world entity across different data sources (e.g., data files, books, websites, and databases). (Source: Wikipedia)

Blocking is a crucial step in any entity resolution pipeline because a pair-wise comparison of all records across two data sources is infeasible. Blocking applies a computationally cheap method to generate a smaller set of candidate record pairs reducing the workload of the matcher. During matching a more expensive pair-wise matcher generates a final set of matching record pairs.

Survey on blocking:

평가 결과

데이터셋과 평가 조건을 선택해 성능·논문·코드를 비교하세요. 보존된 Papers with Code 기록으로, 최신 순위는 아닙니다.

Abt-Buy

원본 태스크: Data Integration / Entity Resolution / Blocking

평가 기록 6건

Amazon-Google

원본 태스크: Data Integration / Entity Resolution / Blocking

평가 기록 6건

WDC Block - large

원본 태스크: Data Integration / Entity Resolution / Blocking

평가 기록 2건

WDC Block - medium

원본 태스크: Data Integration / Entity Resolution / Blocking

평가 기록 2건

WDC Block - small

원본 태스크: Data Integration / Entity Resolution / Blocking

평가 기록 2건

출처: Papers with Code · CC BY-SA 4.0. 아카이브 자료와 새로 수집한 논문은 별도로 관리합니다.