公开(公告)号：US20140229234A1

公开(公告)日：2014-08-14

申请号：US14253138

申请日：2014-04-15

Applicant: Facebook, Inc.

Inventor： Rong Yan ,  John Hegeman

IPC: G06Q30/02 ,  G06N5/04

CPC classification number: G06Q30/0245 ,  G06N5/048 ,  G06Q30/02 ,  G06Q30/0202

Abstract: A social networking system presents content items to users, who then provide feedback regarding pairs of content items. The feedback includes a selection of a content item of the pair of content items that was preferred by the user over the other content item. The social networking system uses this information to train a predictive model that scores content items based on quality. The content items may be advertisements. The social networking system uses the pair-wise comparisons of the advertisements to determine feedback coefficients in an advertising quality score prediction model using regression analysis of the pair-wise comparisons for each predictive factor in the model. In this way, the pair-wise comparisons are used to train the prediction model to understand which advertisements are more enjoyable than others. A feedback coefficient for each predictive factor may be computed based on the preferences received from the group of users.

Abstract translation: 社交网络系统向用户呈现内容，然后他们提供关于内容对的反馈。 反馈包括用户对其他内容项目优选的内容项目对的内容项的选择。 社交网络系统使用该信息来训练基于质量对内容项进行评分的预测模型。 内容项可以是广告。 社交网络系统使用广告的成对比较来确定广告质量得分预测模型中的反馈系数，其使用对于模型中的每个预测因子的成对比较的回归分析。 以这种方式，使用成对比较来训练预测模型，以了解哪些广告比其他广告更愉快。 可以基于从用户组接收的偏好来计算每个预测因子的反馈系数。

公开(公告)号：US09582812B2

公开(公告)日：2017-02-28

申请号：US14253138

申请日：2014-04-15

Applicant: Facebook, Inc.

Inventor： Rong Yan ,  John Hegeman

IPC: G06Q30/02 ,  G06N5/04

CPC classification number: G06Q30/0245 ,  G06N5/048 ,  G06Q30/02 ,  G06Q30/0202

Abstract: A social networking system presents content items to users, who then provide feedback regarding pairs of content items. The feedback includes a selection of a content item of the pair of content items that was preferred by the user over the other content item. The social networking system uses this information to train a predictive model that scores content items based on quality. The content items may be advertisements. The social networking system uses the pair-wise comparisons of the advertisements to determine feedback coefficients in an advertising quality score prediction model using regression analysis of the pair-wise comparisons for each predictive factor in the model. In this way, the pair-wise comparisons are used to train the prediction model to understand which advertisements are more enjoyable than others. A feedback coefficient for each predictive factor may be computed based on the preferences received from the group of users.

Abstract translation: 社交网络系统向用户呈现内容，然后他们提供关于内容对的反馈。 反馈包括用户对其他内容项目优选的内容项目对的内容项的选择。 社交网络系统使用该信息来训练基于质量对内容项进行评分的预测模型。 内容项可以是广告。 社交网络系统使用广告的成对比较来确定广告质量得分预测模型中的反馈系数，其使用对于模型中的每个预测因子的成对比较的回归分析。 以这种方式，使用成对比较来训练预测模型，以了解哪些广告比其他广告更愉快。 可以基于从用户组接收的偏好来计算每个预测因子的反馈系数。

公开(公告)号：US20140222605A1

公开(公告)日：2014-08-07

申请号：US13758553

申请日：2013-02-04

Applicant: Facebook, Inc.

Inventor： Yigal Dan Rubinstein ,  Yanxin Shi ,  John Hegeman

IPC: G06Q30/06 ,  G06Q50/00 ,  G06Q30/08

CPC classification number: G06Q30/0631 ,  G06Q30/08 ,  G06Q50/01

Abstract: A social networking system presents recommendation units to its users. The recommendation units suggest actions for the users to increase their engagement with the social networking system or otherwise interact with other users. The social networking system establishes internal goals and associates bids for recommendation units with different goals. The bids reflect the value to the goal of a user interacting with a recommendation unit. Based on bids for recommendation units associated with one or more goals, expected values of the recommendation units arid determined. The recommendation units are ranked based on the expected values and one or more recommendation units are selected based on the ranking.

Abstract translation: 社交网络系统向其用户提供推荐单位。 推荐单位建议用户采取行动，增加他们与社交网络系统的参与度，或以其他方式与其他用户互动。 社交网络系统建立内部目标，并将不同目标的推荐单位的出价联系起来。 出价反映了与推荐单位交互的用户的目标的价值。 根据与一个或多个目标相关的推荐单位的出价，推荐单位的预期值和确定的值。 基于预期值对推荐单元进行排序，并且基于排名选择一个或多个推荐单位。