Determining temporal relevance of newsfeed stories

Patent metadata

Field	Value
Publication number	US-10063513-B2
Application number	US-201414535308-A
Country	US
Kind code	B2
Filing date	Nov 6, 2014
Priority date	Nov 6, 2014
Publication date	Aug 28, 2018
Grant date	Aug 28, 2018

Abstract

A social networking system generates stories based on actions of users in the system and provides a newsfeed to users that contain stories that related to one or more of their friends in the system. Although the story ranking algorithm includes a time decay to penalize older stories, stories may actually become stale at different rates. To measure the staleness of a story, the system computes a ratio of a current engagement rate for the story to an average engagement rate for the story. Based on this ratio, the system may filter out stale stories, includes the ratio as a feature in the scoring model, and/or adjust the decay rate.

First claim

What is claimed is:

1. A computer-implemented method, comprising:
receiving one or more candidate stories for presentation in newsfeeds for a plurality of users of a social networking system;
identifying a subset of candidate stories from the one or more candidate stories, the identified subset of candidate stories eligible for presentation in a newsfeed for a user of a social networking system; and
applying a filter to the subset of candidate stories to select one or more stories for inclusion in the newsfeed, wherein applying the filter comprises, for each candidate story in the subset of candidate stories:
determining a measure of expected engagement associated with the candidate story based on a rate at which the plurality of users of the social networking system have interacted with the candidate story since the candidate story was available for presentation in the social networking system,
determining a measure of current engagement associated with the candidate story based on the rate at which the plurality of users of the social networking systems have interacted with the candidate story in a current time period,
computing a temporal relevance of the candidate story based on a comparison between the measure of expected engagement and the measure of current engagement, the temporal relevance indicating a likelihood that the candidate story is relevant to the user in the current time period,
determining a decay factor for the candidate story based on the computed temporal relevance, the decay factor indicating a rate at which the candidate story loses relevance,
determining a relevance score for the candidate story based on the determined decay factor,
responsive to determining that the relevance score is greater than a threshold value, selecting the candidate story for presentation in the newsfeed.

2. The method of claim 1 , wherein determining the measure of current engagement comprises:
identifying a plurality of interactions between one or more users of the social networking system and the candidate story;
weighting each of the plurality of interactions based on a type of interaction; and
combining the weighted plurality of interactions to determine the measure of current engagement.

3. The method of claim 1 , wherein the measure of expected engagement is for a given unit of time, and the current time period is the unit of time that most recently elapsed.

4. The method of claim 1 , further comprising computing the temporal relevance based on a ratio of the measure of expected engagement and the measure of current engagement.

5. The method of claim 1 , further comprising computing the temporal relevance based on a difference between the measure of expected engagement and the measure of current engagement.

6. The method of claim 1 , wherein determining whether to select the candidate story comprises computing a relevance score for the candidate story based on a combination of the temporal relevance and one or more other features of the candidate story.

7. A non-transitory computer-readable storage medium containing computer program code for:
receiving one or more candidate stories for presentation in newsfeeds for a plurality of users of a social networking system;
identifying a subset of candidate stories the one or more candidate stories, the identified subset of candidate stories eligible for presentation in a newsfeed for a user of a social networking system; and
applying a filter to the subset of candidate stories to select one or more stories for inclusion in the newsfeed, wherein applying the filter comprises, for each candidate story in the subset of candidate stories:
determining a measure of expected engagement associated with the candidate story based on a rate at which the plurality of users of the social networking system have interacted with the candidate story since the candidate story was available for presentation in the social networking system,
determining a measure of current engagement associated with the candidate story based on a rate at which the plurality of users of the social networking systems have interacted with the candidate story in a current time period,
computing a temporal relevance of the candidate story based on a comparison between the measure of expected engagement and the measure of current engagement, the temporal relevance indicating a likelihood that the candidate story is relevant to the user in the current time period,
determining a decay factor for the candidate story based on the computed temporal relevance, the decay factor indicating a rate at which the candidate story loses relevance,
determining a relevance score for the candidate story based on the determined decay factor, and
responsive to determining that relevance score is greater than a threshold value, selecting the candidate story for presentation in the newsfeed.

8. The non-transitory computer-readable storage medium of claim 7 , wherein code for determining the measure of current engagement comprises code for:
identifying a plurality of interactions between one or more users of the social networking system and the candidate story;
weighting each of the plurality of interactions based on a type of interaction; and
combining the weighted plurality of interactions to determine the measure of current engagement.

9. The non-transitory computer-readable storage medium of claim 7 , wherein the measure of expected engagement is for a given unit of time, and the current time period is the unit of time that most recently elapsed.

10. The non-transitory computer-readable storage medium of claim 7 , further comprising code for computing the temporal relevance based on a ratio of the measure of expected engagement and the measure of current engagement.

11. The non-transitory computer-readable storage medium of claim 7 , further comprising code for computing the temporal relevance based on a difference between the measure of expected engagement and the measure of current engagement.

12. The non-transitory computer-readable storage medium of claim 7 , wherein determining whether to select the candidate story comprises computing a relevance score for the candidate story based on a combination of the temporal relevance and one or more other features of the candidate story.

13. The method of claim 1 , wherein computing a temporal relevance further comprises:
determining a length of time elapsed since the candidate story was first made available to the user;
determining a level of interaction with the candidate story since the candidate story was first made available to the user; and
determining the temporal relevance based on the length of time elapsed since the candidate story was first made available to the user and the level of interaction with the candidate story since the candidate story was first made available to the user.

14. The non-transitory computer readable storage medium of claim 7 , wherein code for computing a temporal relevance comprises code for:
determining a length of time elapsed since the candidate story was first made available to the user;
determining a level of interaction with the candidate story since the candidate story was first made available to the user; and
determining the temporal relevance based on the length of time elapsed since the candidate story was first made available to the user and the level of interaction with the candidate story since the candidate story was first made available to the user.

15. A computer system comprising:
one or more processors;
a non-transitory computer-readable storage medium storing computer program code for:
re

Assignees

• Facebook Inc

Inventors

• Owens Erich James
• Vickrey David

Classifications

• G06Q10/40

  Business processes related to social networking or social networking services · CPC title

• H04L67/22

  Electricity · mapped topic

• G06Q50/01

  Physics · mapped topic

• H04L51/32 Primary

  Electricity · mapped topic

• G06F17/30867

  Physics · mapped topic