公开(公告)号：US09424340B1

公开(公告)日：2016-08-23

申请号：US14521078

申请日：2014-10-22

Applicant: Google Inc.

Inventor： Rupesh Kapoor ,  David Michael Proudfoot ,  Joachim Kupke

IPC: G06F17/30

CPC classification number: G06F17/30613 ,  G06F17/3053 ,  G06F17/3071 ,  G06F17/30864

Abstract: A system may identify a set of first documents associated with an organization, and identify clusters to which the first documents belong. Each of a number of the identified clusters may include a group of documents that includes one of the first documents and one or more second documents associated with one or more different organizations. The system may determine a quality score for each of the documents in each of the identified clusters, and determine, for each of the number of the identified clusters, whether the quality score of the one of the first documents in the identified cluster is higher than the quality score of the one or more second documents in the identified cluster. The system may generate a proxy pad score based on the determinations, and store the proxy pad score.

Abstract translation: 系统可以标识与组织相关联的一组第一文档，并且识别第一文档所属的群集。 多个所识别的集群中的每一个可以包括一组文档，其包括第一文档之一和与一个或多个不同组织相关联的一个或多个第二文档。 所述系统可以确定每个所识别的集群中的每个文档的质量得分，并且对于所识别的集群中的每一个，确定所识别的集群中的所述第一文档之一的质量得分是否高于 所识别的群集中的一个或多个第二个文档的质量得分。 该系统可以基于确定产生代理贴片分数，并存储代理贴片分数。