Estimating social interest in time-based media
Social media content items are mapped to relevant time-based media events. These mappings may be used as the basis for multiple applications, such as ranking of search results for time-based media, automatic recommendations for time-based media, prediction of audience interest for media purchasing/planning, and estimating social interest in the time-based media. Social interest in time-based media (e.g., video and audio streams and recordings) segments is estimated through a process of data ingestion and integration. The estimation process determines social interest in specific events represented as segments in time-based media, such as particular plays in a sporting event, scenes in a television show, or advertisements in an advertising block. The resulting estimates of social interest also can be graphically displayed.
1. A computer-executed method for associating social media content items with a time-based media event, the method comprising:
accessing from a social networking system a plurality of candidate social media content items authored by users of the social networking system;
for each of the candidate social media content items, determining a confidence score indicative of a probability that the candidate social media content item is relevant to the event;
aligning with the event, based on their respective confidence scores, a subset of the plurality of the social media content items; and
collecting in a data store the alignments between the event and the subset of the plurality of the social media content items.
2. The computer-executed method of claim 1 , further comprising segmenting a received time-based medium into a plurality of segments corresponding to time-based media events.
3. The computer-executed method of claim 1 , wherein determining the confidence score indicative of the probability that the candidate social media content item is relevant to the event further comprises:
extracting event features from annotations associated with the event;
extracting social media features from the plurality of social media content items; and
mapping the event to the social media content items based on a relationship between the event features and social media features.
4. The computer-executed method of claim 3 , further comprising annotating the event with the annotations using metadata instances relevant to the event.
5. The computer-executed method of claim 1 , further comprising:
aggregating the confidence scores of the associated subset of candidate social media content items to produce an aggregate score; and
determining a level of social interest in the event based upon the aggregate score.
6. A computer-executed method for determining social interest in an event in a time-based media, the method comprising:
accessing from a social networking system a plurality of candidate social media content items authored by users of the social networking system;
for each of the candidate social media content items, determining a confidence score indicative of a probability that the candidate social media content item is relevant to the event;
aligning with the event, based on their respective confidence scores, a subset of the plurality of the social media content items;
aggregating the confidence scores of the aligned subset of candidate social media content items to produce an aggregate score; and
determining a level of social interest in the event based upon the aggregate score.
7. The computer-executed method of claim 6 , wherein determining the confidence score indicative of the probability that the candidate social media content item is relevant to the event further comprises:
extracting event features from annotations associated with the event;
extracting social media features from the plurality of social media content items; and
mapping the event to the social media content items based on a relationship between the event features and social media features.
8. The computer-executed method of claim 7 , further comprising annotating the event with the annotations using metadata instances relevant to the event.
9. A computer-executed method for determining social interest in time-based media events, the method comprising:
selecting a plurality of events in a time-based medium;
accessing from a social networking system a plurality of candidate social media content items authored by users of the social networking system that are potentially relevant to the events in the time based medium;
for each social media content item, determining a confidence score for the social media content item with respect to at least one of the plurality of events, the confidence score indicating a probability that the social media content item is relevant to the at least one event; and
for each of the plurality of events:
aligning with the event, based on their respective confidence scores, a subset of the plurality of the social media content items;
aggregating the confidence scores of the social media content items having a confidence score for the event; and
determining a level of social interest for in the event based on the aggregate score.
10. The computer-executed method of claim 9 , further comprising collecting in a data store the alignments between the event and the subset of the plurality of the social media content items.
11. The computer-executed method of claim 9 , wherein for a selected event, determining the confidence score indicative of the probability that the candidate social media content item is relevant to the event further comprises:
extracting event features from annotations associated with the event;
extracting social media features from the plurality of social media content items; and
mapping the event to the social media content items based on a relationship between the event features and social media features.
12. The computer-executed method of claim 11 , further comprising annotating the event with the annotations using metadata instances relevant to the event.
13. A computer-executed method for mapping social content items to time-based media events, comprising:
accessing from a social networking system a plurality of social media content items authored by users of the social networking system;
aligning a plurality of metadata instances to segments of time-based media corresponding to events in the time-based media to form annotated events; and
mapping to the annotated events the social media content items relevant to the annotated events.
14. The computer-executed method of claim 13 , wherein aligning a plurality of metadata instances to segments of time-based media corresponding to events in the time-based media to form annotated events further comprises:
receiving the plurality of metadata instances;
segmenting the time-based media into the segments corresponding to events in the time-based media, each segment having a beginning and an end; and
determining, for each metadata instance, a segment of the time-based media that most likely aligns with the metadata instance.
15. The computer-executed method of claim 13 , wherein mapping to the annotated events a plurality of social media content items relevant to the annotated events further comprises:
extracting social media features from the plurality of social media content items;
extracting event features from the annotated events; and
mapping the annotated events to the plurality of social media content items based on a relationship between the event features and social media features.
16. A system for associating social media content items with a time-based media event, the system comprising:
means for accessing from a social networking system a plurality of candidate social media content items authored by users of the social networking system;
means for determining a confidence score for each of the candidate social media content items indicative of a probability that the candidate social media content item is relevant to the event;
means for aligning with the event, based on their respective confidence scores, a subset of the plurality of the social media content items; and
a data store for collecting the alignments between the event and the subset of the plurality of the social media content items.
17. The system of claim 16 , further comprising:
means for extracting event features from annotations associated with the event;
means for extracting social media features from the plurality of social media content items; and
means for mapping the event to the social media content items based on a relationship between the event features and social media features.
18. The system of claim 17 , further comprising annotating the event with the annotations using metadata instances relevant to the event.
19. The system of claim 16 , further comprising:
means for aggregating the confidence scores of the associated subset of candidate social media content items to produce an aggregate score; and
means for determining a level of social interest in the event based upon the aggregate score.
20. A system for mapping social content items to time-based media events, comprising:
means for accessing from a social networking system a plurality of social media content items authored by users of the social networking system;
means for aligning a plurality of metadata instances to segments of time-based media corresponding to events in the time-based media to form annotated events; and
means for mapping to the annotated events the social media content items relevant to the annotated events.
21. A system for associating social media content items with a time-based media event, the system comprising:
a computer processor; and
a computer-readable storage medium storing computer program modules configured to execute on the computer processor, the computer program modules comprising:
a data ingestion engine configured to access from a social networking system a plurality of candidate social media content items authored by users of the social networking system;
a media/event alignment engine configured to determine a confidence score for each of the candidate social media content items indicative of a probability that the candidate social media content item is relevant to the event and to align with the event, based on their respective confidence scores, a subset of the plurality of the social media content items; and
a data store for collecting the alignments between the event and the subset of the plurality of the social media content items.
22. The system of claim 21 , the computer program modules further comprising:
a comparative feature extraction engine configured to extract event features from annotations associated with the event and to extract social media features from the plurality of social media content items; and
the media/event alignment engine further configured to map the event to the social media content items based on a relationship between the event features and social media features.
23. The system of claim 22 , the computer program modules further comprising:
an annotation engine configured to annotate the event with the annotations using metadata instances relevant to the event.
24. The system of claim 21 , the computer program modules further comprising:
a social interest estimator configured to aggregate the confidence scores of the associated subset of candidate social media content items to produce an aggregate score and to determine a level of social interest in the event based upon the aggregate score.
25. A computer system for mapping social content items to time-based media events, the computer system comprising a processor and a memory and further comprising:
a data ingestion engine, executed by the computer system, configured to access from a social networking system a plurality of candidate social media content items authored by users of the social networking system;
a metadata alignment engine, executed by the computer system, configured to align a plurality of metadata instances to segments of time-based media corresponding to events in the time-based media to form annotated events; and
a media/event alignment engine, executed by the computer system, configured to map to the annotated events a plurality of social media content items relevant to the annotated events.