Ranking feed based on a likelihood of a user posting original broadcast story in a session

A non-transitory computer-readable storage medium comprising stored instructions for:

• receiving, by a social networking system, a request for feed from a client device of a target user, wherein a feed represents a sequence feed stories, each feed story comprising one or more of text, image, or video;
• selecting a feed to provide to the user in response to the request, the feed comprising a plurality of feed stories comprising original broadcast feed stories and non-original broadcast feed stories;
• accessing a model configured to receive an input feature vector describing the request received from the client device and determine an output indicating a likelihood of the user posting an original broadcast feed story to a set of users of the social networking system;
• determining a feature vector describing the request received from the client device, the feature vector comprising a feature specifying a time of day of the request;
• providing the feature vector as input to the model;
• determining based on the model, whether a likelihood of the user posting an original broadcast feed story to a set of users of the social networking system is above a threshold value;
• responsive to determining that the likelihood of the user posting the original broadcast feed story is greater than a threshold value, ranking the feed stories of the feed for presenting to the user such that original broadcast feed stories are ranked higher in the feed compared to a corresponding feed presented to the user responsive to the likelihood of the user posting the original broadcast feed story being less than the threshold value;
• sending the plurality of feed stories to the client device of the target user.

A computer system comprising:

• a computer processor;
• and
• a non-transitory computer-readable storage medium comprising stored instructions for:
• receiving, by a social networking system, a request for feed from a client device of a target user, wherein a feed represents a sequence feed stories, each feed story comprising one or more of text, image, or video;
• selecting a feed to provide to the user in response to the request, the feed comprising a plurality of feed stories comprising original broadcast feed stories and non-original broadcast feed stories;
• accessing a model configured to receive an input feature vector describing the request received from the client device and determine an output indicating a likelihood of the user posting an original broadcast feed story to a set of users of the social networking system;
• determining a feature vector describing the request received from the client device, the feature vector comprising a feature specifying a time of day of the request;
• providing the feature vector as input to the model;
• determining based on the model, whether a likelihood of the user posting an original broadcast feed story to a set of users of the social networking system is above a threshold value;
• responsive to determining that the likelihood of the user posting the original broadcast feed story is greater than a threshold value, ranking the feed stories of the feed for presenting to the user such that original broadcast feed stories are ranked higher in the feed compared to a corresponding feed presented to the user responsive to the likelihood of the user posting the original broadcast feed story being less than the threshold value;
• sending the plurality of feed stories to the client device of the target user.