User interaction based related videos

• US 9,088,808 B1
• Filed: 02/08/2008
• Issued: 07/21/2015
• Est. Priority Date: 02/08/2008
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: He, Yu, Stoutamire, David Petrie
• Primary Examiner(s): Bruckart, Benjamin R
• Assistant Examiner(s): Thomas, Jason
• Application Number:
  US12/028,710
• Time in Patent Office: 2,720 Days
• Field of Search: 707/736, 707/748, 725/38, 725/44, 725/46
• US Class Current: 1/1
• CPC Class Codes: H04N 21/251 Learning process for intelligent management, e.g. learning user preferences for recommending movies details of learning user preferences for the retrieval of video data in a video database G06F16/739; computer systems using learning methods G06N3/08
• CPC Class Codes: H04N 21/252 Processing of multiple end-users' preferences to derive collaborative data
• CPC Class Codes: H04N 21/44204 Monitoring of content usage, e.g. the number of times a movie has been viewed, copied or the amount which has been watched monitoring of user activities for profile generation for accessing a video database G06F16/739; protecting generic digital content where the protection is independent of the precise nature of the content G06F21/10; arrangements for monitoring the use made of the broadcast services in broadcast systems H04H60/31
• CPC Class Codes: H04N 21/4662 characterized by learning algorithms
• CPC Class Codes: H04N 21/4667 Processing of monitored end-user data, e.g. trend analysis based on the log file of viewer selections
• CPC Class Codes: H04N 21/4756 for rating content, e.g. scoring a recommended movie

Abstract

Methods and systems to locate related videos in a video access log. In one embodiment, the method comprises extracting user events from a log database, tagging each event as positive or negative, determining if a video is positively interacted and processing the tagged videos in a sliding window to determine positive interactions between a pair of videos.

Claims

1. A computer implemented method for selecting a set of videos related to a first video, the method comprising:

receiving data representative of interactions with videos by a plurality of users;

identifying, from the received data for a first video, those interactions that are positive interactions with the first video, the positive interactions indicating user interest in the first video that occurred after playback of the first video began;

identifying, from the received data for a plurality of second videos, those interactions that are positive interactions with the plurality of second videos, the positive interactions indicating user interest in the plurality of second videos that occurred after playback of the plurality of second videos began, wherein playback of the plurality of second videos occurred within a locality of time after playback of the first video;

generating co-occurring positive interactions for the first video and each of the plurality of second videos based on the identified positive interactions, a co-occurring positive interaction for the first video and a second video indicating interest of a user in both the first video and the second video after playback of the first video and playback of the second video within the locality of time;

assigning a weight to each of a plurality of the co-occurring positive interactions for the first video and each of the plurality of second videos, the assigned weight for a co-occurring positive interaction for the first video and each of the plurality of second videos based on a length of time between viewing of the first video and each of the plurality of second videos;

storing a count of co-occurring positive interactions for the first video and each of the plurality of second videos based on the weighted co-occurring positive interactions for the first video and each of the plurality of second videos;

ranking, by a computer, each of the plurality of second videos using the counts of co-occurring positive interactions;

selecting the set of related videos for the first video from the plurality of second videos using the ranked counts of co-occurring positive interactions;

ranking, by the computer, the plurality of second videos included in the set of related videos for the first video based at least in part on a total duration of watch time of each of the plurality of second videos by the plurality of users; and

storing an indication of the set of related videos.

2. The computer implemented method of claim 1, further comprising:

storing for each of the plurality of users, user events indicating positive interactions with the first video; and

storing for each of the plurality of users, user events indicating positive interactions with any of the plurality of second videos.

3. The computer implemented method of claim 2, further comprising:

determining from the user events, a count of the co-occurring positive interactions by individual users of the first video and each of the second videos.

4. The computer implemented method of claim 2, wherein user events indicating positive interactions comprises a video watch ratio greater than a predetermined value.

5. The computer implemented method of claim 1, wherein storing the indication of the set of related videos comprises:

generating a list of related videos for the first video using the selected set of related videos from the ranked plurality of second videos.

6. The computer implemented method of claim 1, further comprising:

filtering the count of co-occurring positive interactions to remove interactions indicating that the first video and one of the plurality of second videos were accessed by a malicious software program that accessed the first video and the plurality of second videos.

17. The computer implemented method of claim 1, wherein the counted co-occurring positive interactions occur over the locality of time.

7. A computer implemented method for determining co-occurring positive interactions between videos to create a list of related videos, the method comprising:

receiving data representative of interactions with videos by a plurality of users;

for at least one video, identifying from the received data those interactions that are positive interactions with the at least one video, the positive interactions indicating user interest in the at least one video that occurred after the at least one video loaded for display;

for a plurality of second videos, identifying from the received data those interactions that are positive interactions with the plurality of second videos, the positive interactions indicating user interest in the plurality of second videos that occurred after playback of the plurality of second videos began, wherein playback of the plurality of second videos occurred within a locality of time after playback of the first video;

for a plurality of users, determining, by a computer system, co-occurring positive interactions for pairs of videos interacted with by the plurality of users, wherein a co-occurring positive interaction for a pair of videos indicates user interest in both the at least one video and a second video from the plurality of second videos after the at least one video loaded for display and the second video loaded for display within the locality of time;

for the at least one video, assigning a weight to each of a plurality of the co-occurring positive interactions for the at least one video and each of the plurality of second videos included in the pairs of videos, the assigned weight for a co-occurring positive interaction for the at least one video and each of the plurality of second videos included in the pairs of videos based on a length of time between viewing of the at least one video and each of the plurality of second videos included in the pairs of videos;

for the at least one video, determining counts of co-occurring positive interactions between the at least one video and each of the plurality of second videos included in the pairs of videos based on the weighted co-occurring positive interactions for the at least one video and each of the plurality of second videos included in the pairs of videos;

for the at least one video, ranking, by a computer, each of the plurality of second videos using the counts of co-occurring positive interactions between the at least one video and the plurality of second videos;

for the at least one video, selecting a list of related videos for the at least one video from the plurality of second videos using the ranked counts of co-occurring positive interactions; and

for the at least one video, ranking, by the computer, the plurality of second videos included in the list of related videos based at least in part on a total duration of watch time of each of the plurality of second videos by the plurality of users.

8. A video hosting website for determining related videos, the video hosting website comprising:

a video database storing videos;

a video access log comprising events, wherein the events are received data representative of user interactions with videos from a plurality of users, the events including positive interactions;

a user interaction analysis server configured to;

identify, from the received data for a first video, those interactions that are positive interactions with the first video, the positive interactions indicating user interest in the first video that occurred after playback of the first video began;

identify, from the received data for a plurality of second videos, those interactions that are positive interactions with the plurality of second videos, the positive interactions indicating user interest in the plurality of second videos that occurred after playback of the plurality of second videos began, wherein playback of the plurality of second videos occurred within a locality of time after playback of the first video;

generate co-occurring positive interactions for the first video and each of the plurality of second videos based on the identified positive interactions, a co-occurring positive interaction for the first video and a second video indicating interest of a user in both the first video and the second video after playback of the first video and playback of the second video within the locality of time;

assign a weight to each of a plurality of the co-occurring positive interactions for the first video and each of the plurality of second videos, the assigned weight for a co-occurring positive interaction for the first video and each of the plurality of second videos based on a length of time between viewing of the first video and each of the plurality of second videos;

determine counts of co-occurring positive interactions for the first video and each of the plurality of second videos based on the weighted co-occurring positive interactions for the first video and each of the plurality of second videos;

rank each of the plurality of second videos using the counts of co-occurring positive interactions;

select the set of related videos for the first video from the plurality of second videos using the ranked counts of co-occurring positive interactions; and

rank the plurality of second videos included in the set of related videos for the first video based at least in part on a total duration of watch time of each of the plurality of second videos by the plurality of users; and

a related videos database adapted to store for the first video an indication of the set of related videos.

9. A computer program product for selecting a set of videos related to a first video, comprising a non-transitory computer-readable storage medium containing executable computer program code for:

receiving data representative of interactions with videos by a plurality of users;

identifying, from the received data for a first video, those interactions that are positive interactions with the first video, the positive interactions indicating user interest in the first video that occurred after playback of the first video began;

identifying, from the received data for a plurality of second videos, those interactions that are positive interactions with the plurality of second videos, the positive interactions indicating user interest in the plurality of second videos that occurred after playback of the plurality of second videos began, wherein playback of the plurality of second videos occurred within a locality of time after playback of the first video;

generating co-occurring positive interactions for the first video and each of the plurality of second videos based on the identified positive interactions, a co-occurring positive interaction for the first video and a second video indicating interest of a user in both the first video and the second video after playback of the first video and playback of the second video within the locality of time;

assigning a weight to each of a plurality of the co-occurring positive interactions for the first video and each of the plurality of second videos, the assigned weight for a co-occurring positive interaction for the first video and each of the plurality of second videos based on a length of time between viewing of the first video and each of the plurality of second videos;

storing a count of co-occurring positive interactions for the first video and each of the plurality of second videos based on the weighted co-occurring positive interactions for the first video and each of the plurality of second videos;

ranking, by a computer, each of the plurality of second videos using the counts of co-occurring positive interactions;

selecting the set of related videos for the first video from the plurality of second videos using the ranked counts of co-occurring positive interactions;

ranking, by the computer, the plurality of second videos included in the set of related videos for the first video based at least in part on a total duration of watch time of each of the plurality of second videos by the plurality of users; and

storing an indication of the set of related videos.

10. The computer program product of claim 9, the method further comprising:

storing for each of the plurality of users, user events indicating positive interactions with the first video; and

storing for each of the plurality of users, user events indicating positive interactions with any of the plurality of second videos.

11. The computer program product of claim 10, wherein user events indicating positive interactions comprises a video watch ratio greater than a predetermined value.

12. The computer program product of claim 10, the method further comprising:

determining from the user events, a count of the co-occurring positive interactions by individual users of the first video and each of the second videos.

13. The computer program product of claim 9, wherein storing the indication of the set of related videos comprises:

generating a list of related videos for the first video using the selected set of related videos from the ranked plurality of second videos.

14. The computer program product of claim 9, the method further comprising:

filtering the count of co-occurring positive interactions to remove interactions indicating that the first video and one of the plurality of second videos were accessed by a software program, the software program not operated by a user.

15. A computer program product for determining co-occurring positive interactions between videos to create a list of related videos comprising, a non-transitory computer readable storage medium storing computer executable instructions for:

receiving data representative of interactions with videos by a plurality of users;

for at least one video, identifying from the received data those interactions that are positive interactions with the at least one video, the positive interactions indicating user interest in the at least one video that occurred after the at least one video loaded for display;

for a plurality of second videos, identifying from the received data those interactions that are positive interactions with the plurality of second videos, the positive interactions indicating user interest in the plurality of second videos that occurred after playback of the plurality of second videos began, wherein playback of the plurality of second videos occurred within a locality of time after playback of the first video;

for a plurality of users, determining, by a computer system, co-occurring positive interactions for pairs of videos interacted with by the plurality of users, wherein a co-occurring positive interaction for a pair of videos indicates user interest in both the at least one video and a second video from the plurality of second videos after the at least one video loaded for display and the second video loaded for display within the locality of time;

for the at least one video, assigning a weight to each of a plurality of the co-occurring positive interactions for the at least one video and each of the plurality of second videos included in the pairs of videos, the assigned weight for a co-occurring positive interaction for the at least one video and each of the plurality of second videos included in the pairs of videos based on a length of time between viewing of the at least one video and each of the plurality of second videos included in the pairs of videos;

for the at least one video, determining counts of co-occurring positive interactions between the at least one video and each of the plurality of second videos included in the pairs of videos based on the weighted co-occurring positive interactions for the at least one video and each of the plurality of second videos included in the pairs of videos;

for the at least one video, ranking, by a computer, each of the plurality of second videos using the counts of co-occurring positive interactions between the at least one video and the plurality of second videos;

for the at least one video, selecting a list of related videos for the at least one video from the plurality of second videos ranked using the ranked counts of co-occurring positive interactions; and

for the at least one video, ranking, by the computer, the plurality of second videos included in the list of related videos based at least in part on a total duration of watch time of each of the plurality of second videos by the plurality of users.

16. A computer program product for identifying related videos in plurality of videos V comprising video (V₁, . . . V_n), comprising a non-transitory computer-readable storage medium containing executable computer program code for:

for at least one pair of videos (V_i, V_j) in the plurality of videos V, generating co-occurring positive interactions for the at least one pair of videos (V_i, V_j) in the plurality of videos V wherein a co-occurring positive interaction for a pair of videos (V_i, V_j) indicates interest of a user in both the video V_i and the video V_j that occurred after the video V_i loaded for display and the video V_j loaded for display, wherein the video V_j loaded for display within a locality of time that video V_i loaded for display;

for at least one pair of videos (V_i, V_j) in the plurality of videos V, assigning a weight to each of a plurality of the co-occurring positive interactions for the at least one pair of videos (V_i, V_j), the assigned weight for a co-occurring positive interaction for the at least one pair of videos (V_i, V_j) based on a length of time between viewing video V_i and each of the second videos V_j;

for at least one pair of videos (V_i, V_j) in the plurality of videos V, storing a count C_i,j the co-occurring positive user interactions for the pair of videos (V_i, V_j) based on the weighted co-occurring positive interactions for the pair of videos (V_i, V_j) for at least one video V_i, ranking the videos V_j using the counts C_i,j;

for at least one video V_i, identifying a set R_i of related videos V from the ranked second videos V_j, using the ranked counts C_i,j; and

for at least one video V_i, ranking the second videos V_j included in the set R_i of related videos V based at least in part on a total duration of watch time of each of the second videos V_j by a plurality of users that viewed the second videos.

18. A computer implemented method for selecting a set of videos related to a first video, the method comprising:

receiving data representative of interactions with videos by a plurality of users;

identifying, from the received data for a first video, those interactions that are positive interactions with the first video, the positive interactions indicating user interest in the first video that occurred after playback of the first video began;

identifying, from the received data for a plurality of second videos, those interactions that are positive interactions with the plurality of second videos, the positive interactions indicating user interest in the plurality of second videos that occurred after playback of the plurality of second videos began, wherein playback of the plurality of second videos occurred within a locality of time after playback of the first video;

generating co-occurring positive interactions for the first video and each of the plurality of second videos based on the identified positive interactions, a co-occurring positive interaction for the first video and a second video indicating interest of a user in both the first video and the second video after playback of the first video and playback of the second video within the locality of time;

assigning a weight to each of a plurality of the co-occurring positive interactions for the first video and each of the plurality of second videos, the assigned weight for a co-occurring positive interaction for the first video and each of the plurality of second videos based on a length of time between viewing of the first video and each of the plurality of second videos;

storing a count of co-occurring positive interactions for the first video and each of the plurality of second videos based on the weighted co-occurring positive interactions for the first video and each of the plurality of second videos;

ranking, by a computer, each of the plurality of second videos based at least in part on the counts of co-occurring positive interactions and a total duration of watch time of each of the plurality of second videos by the plurality of users;

selecting the set of related videos for the first video based at least in part on the ranked plurality of second videos; and

storing an indication of the set of related videos.