公开(公告)号：US10783557B2

公开(公告)日：2020-09-22

申请号：US15331083

申请日：2016-10-21

申请人： Yahoo Holdings, Inc.

发明人： Jonathan Malkin ,  Csaba Kecskemeti

IPC分类号： G06Q30/02

摘要： An electronic system may be configured receive a set of location-based events and form the location-based events into a plurality of groups. The location-based events identify user devices that transmitted messages over the network and physical locations of the user devices when the messages were generated. The groups may be formed based on the physical locations and the number of groups to be formed. In some embodiments, the system may use k-means clustering to form the groups. The system may analyze the groups to determine a home location of at least some of the user devices identified in the set.

公开(公告)号：US20180025086A1

公开(公告)日：2018-01-25

申请号：US15215090

申请日：2016-07-20

申请人： Yahoo Holdings, Inc.

发明人： Jonathan Malkin ,  Jinhui Liu ,  Jerry Zhang ,  Csaba Kecskemeti ,  Wendell Baker

IPC分类号： G06F17/30

CPC分类号： G06F17/30867 ,  G06F17/30861 ,  G06Q30/02 ,  H04L67/306

摘要： An electronic system may be configured to determine if a first-type device ID and a second-type device ID both identify a same device. To do so, the electronic system may preliminarily group a particular first-type ID with a particular second-type ID based on matching time slots and network addresses. Individual scores for each of the preliminary groups may be determined based on cardinalities associated with the time slots. Combined scores may then be determined for groups having the same first-type device ID and second-type device ID. A final analysis may be performed to confirm, with a sufficient level of confidence, whether first-type device IDs and second-type device IDs included in the preliminary groups identify the same device.

公开(公告)号：US20180114249A1

公开(公告)日：2018-04-26

申请号：US15331083

申请日：2016-10-21

申请人： Yahoo Holdings, Inc.

发明人： Jonathan Malkin ,  Csaba Kecskemeti

IPC分类号： G06Q30/02 ,  G06N5/04

摘要： An electronic system may be configured receive a set of location-based events and form the location-based events into a plurality of groups. The location-based events identify user devices that transmitted messages over the network and physical locations of the user devices when the messages were generated. The groups may be formed based on the physical locations and the number of groups to be formed. In some embodiments, the system may use k-means clustering to form the groups. The system may analyze the groups to determine a home location of at least some of the user devices identified in the set.