公开(公告)号：US09378465B2

公开(公告)日：2016-06-28

申请号：US13872811

申请日：2013-04-29

Applicant: Facebook, Inc.

Inventor： Allan Stewart ,  Eugene Zarakhovsky ,  Christopher Palow ,  Chetan Gowda ,  Brent Dorman

IPC: G06N5/02 ,  G06N99/00

CPC classification number: G06N99/005

Abstract: A method of operation of a URL spam detection system includes: identifying a feature dimension of a user action on a social networking system to detect anomalies; extracting URL chunks from a content associated with the user action; aggregating a non-content feature of the user action along the feature dimension into a URL distribution store to produce a feature distribution for each of the URL chunks; determining whether the feature distribution of a particular URL chunk within the URL chunks exceeds an expectation threshold for the feature dimension; and classifying the particular URL chunk as an illegitimate URL when the feature distribution exceeds the expectation threshold to restrict access to a particular URL chunk on a social networking system.

Abstract translation: URL垃圾邮件检测系统的操作方法包括：识别社交网络系统上的用户动作的特征维度以检测异常; 从与用户操作相关联的内容中提取URL块; 将沿着所述特征维度的所述用户动作的非内容特征聚合到URL分发存储器中，以为每个所述URL块生成特征分布; 确定所述URL块内的特定URL块的特征分布是否超过所述特征维度的期望阈值; 并且当特征分布超过期望阈值以限制对社交网络系统上的特定URL块的访问时，将特定URL块分类为非法URL。

公开(公告)号：US20150339705A1

公开(公告)日：2015-11-26

申请号：US14285289

申请日：2014-05-22

Applicant: Facebook, Inc.

Inventor： Vijaye Ganesh Raji ,  Jason Duane Clark ,  Eugene Zarakhovsky ,  Jonathan J. Gross ,  Brent Dorman

IPC: G06Q30/02 ,  G06F21/60

CPC classification number: G06Q30/0248 ,  G06F21/552 ,  G06F21/60 ,  G06Q30/02

Abstract: Embodiments are disclosed for identifying a suspect application based on multiple operating factors from use of multiple applications. The embodiments can generate a representative distribution of a selected factor based on collected information corresponding to multiple operating factors from use of multiple applications. The embodiments can compare a representative distribution of a target factor with the representative distribution of the selected factor and identify a suspect application when these distributions are different.

Abstract translation: 公开了用于基于来自多个应用的​​使用的多个操作因素识别可疑应用的实施例。 基于来自使用多个应用的​​多个操作因素的收集的信息，实施例可以生成所选因子的代表性分布。 实施例可以将目标因子的代表性分布与所选因子的代表性分布进行比较，并且当这些分布不同时识别可疑应用。

公开(公告)号：US11023923B2

公开(公告)日：2021-06-01

申请号：US14285289

申请日：2014-05-22

Applicant: Facebook, Inc.

Inventor： Vijaye Ganesh Raji ,  Jason Duane Clark ,  Eugene Zarakhovsky ,  Jonathan J. Gross ,  Brent Dorman

IPC: G06Q30/02 ,  G06F21/60 ,  G06F21/55

Abstract: Embodiments are disclosed for identifying a suspect application based on multiple operating factors from use of multiple applications. The embodiments can generate a representative distribution of a selected factor based on collected information corresponding to multiple operating factors from use of multiple applications. The embodiments can compare a representative distribution of a target factor with the representative distribution of the selected factor and identify a suspect application when these distributions are different.

公开(公告)号：US20140324741A1

公开(公告)日：2014-10-30

申请号：US13872811

申请日：2013-04-29

Applicant: Facebook, Inc.

Inventor： Allan Stewart ,  Eugene Zarakhovsky ,  Christopher Palow ,  Chetan Gowda ,  Brent Dorman

IPC: G06N5/02 ,  G06N99/00

CPC classification number: G06N99/005

Abstract: A method of operation of a URL spam detection system includes: identifying a feature dimension of a user action on a social networking system to detect anomalies; extracting URL chunks from a content associated with the user action; aggregating a non-content feature of the user action along the feature dimension into a URL distribution store to produce a feature distribution for each of the URL chunks; determining whether the feature distribution of a particular URL chunk within the URL chunks exceeds an expectation threshold for the feature dimension; and classifying the particular URL chunk as an illegitimate URL when the feature distribution exceeds the expectation threshold to restrict access to a particular URL chunk on a social networking system.

Abstract translation: URL垃圾邮件检测系统的操作方法包括：识别社交网络系统上的用户动作的特征维度以检测异常; 从与用户操作相关联的内容中提取URL块; 将沿着所述特征维度的所述用户动作的非内容特征聚合到URL分发存储器中，以为每个所述URL块生成特征分布; 确定所述URL块内的特定URL块的特征分布是否超过所述特征维度的期望阈值; 并且当特征分布超过期望阈值以限制对社交网络系统上的特定URL块的访问时，将特定URL块分类为非法URL。