Information propagation probability for a social network

• US 9,256,829 B2
• Filed: 01/19/2015
• Issued: 02/09/2016
• Est. Priority Date: 12/17/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Zaman, Tauhid Rashed, Van Gael, Jurgen Anne Francois Marie, Stern, David, Herbrich, Ralf, Lotan, Gilad
• Primary Examiner(s): Holmes, Michael B
• Application Number:
  US14/599,995
• Publication Number:
  US 20150134579A1
• Time in Patent Office: 386 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: H04L 43/08 Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
• CPC Class Codes: H04L 51/52 for supporting social networking services

Abstract

One or more techniques and/or systems are disclosed for predicting propagation of a message on a social network. A predictive model is trained to determine a probability of propagation of information on the social network using both positive and negative information propagation feedback, which may be collected while monitoring the social network over a desired period of time for information propagation. A particular message can be input to the predictive model, which can determine a probability of propagation of the message on the social network, such as how many connections may receive at least a portion of the message and/or a likelihood of at least a portion of the message reaching respective connections in the social network.

Claims

1. A method performed using at least one computer-based processor, the method comprising:

monitoring one or more messages on a social network for information propagation feedback, the monitoring comprising collecting propagation metadata and non-propagation metadata;

training a predictive model using the propagation metadata and the non-propagation metadata as a training set;

determining a probability of propagation of an input message on the social network using the predictive model;

retrieving one or more other probabilities for one or more of alternate content for the input message or an alternate time for posting the input message to the social network; and

determining message adjustment information for the input message based at least upon the probability of propagation and the one or more other probabilities, the message adjustment information comprising one or more of a content adjustment for the input message or a timing adjustment for the posting of the input message, wherein the message adjustment information causes the input message to have an increased probability of propagation on the social network.

2. The method of claim 1, comprising collecting an indication of a propagation of a first message of a first user by a second user as at least part of the propagation metadata.

3. The method of claim 1, comprising collecting an indication of a lack of propagation of a first message of a first user by a second user as at least part of the non-propagation metadata.

4. The method of claim 1, comprising collecting an indication of a lack of propagation of a first message of a first user by any connections of the first user as at least part of the non-propagation metadata.

5. The method of claim 1, wherein the non-propagation metadata is associated with negative feedback instances where users of the social network did not propagate individual messages over the social network using a wired or wireless electronic communication network.

6. The method of claim 5, wherein the propagation metadata is associated with positive feedback instances where the users of the social network did propagate other individual messages over the social network using a wired or wireless electronic communication network.

7. The method of claim 6, wherein the positive feedback instances involve the users forwarding or reposting the other individual messages to other users.

8. A system, comprising:

one or more processing units; and

memory comprising instructions that, when executed by at least one of the one or more processing units, cause the at least one of the one or more processing units to;

monitor messages communicated over a wired or wireless electronic network to determine whether the messages are propagated among various computing devices, the messages being monitored by collecting propagation metadata for instances where individual messages are propagated and non-propagation metadata for other instances where other individual messages are not propagated;

train a predictive model using the propagation metadata and the non-propagation metadata as a training set; and

determine a probability of propagation of an input message on the wired or wireless electronic network using the predictive model.

9. The system of claim 8, wherein the instructions, when executed by the at least one of the one or more processing units, cause the at least one of the one or more processing units to:

determine message adjustment information for the input message based on the probability of propagation, the message adjustment information comprising one or more of a content adjustment for the input message or a timing adjustment for the input message.

10. The system of claim 8, wherein users of the various computing devices communicate the messages over a social network and the probability of propagation reflects the likelihood that an individual user of the social network will share the input message with at least one other user of the social network.

11. The system of claim 10, wherein the messages are communicated among the various computing devices over the wired or wireless electronic network using Internet Protocol.

12. The system of claim 11, wherein the propagation metadata identifies Internet Protocol addresses of individual computing devices that are used to propagate the individual messages.

13. The system of claim 8, wherein the predictive model is a Bayesian classifier, a support vector machine, a regression classifier, or a nearest-neighbor classifier.

14. The system of claim 8, wherein the propagation metadata identifies themes, topics, or keywords of the individual messages that are propagated, the non-propagation metadata identifies themes, topics, or keywords of the other individual messages that are not propagated, and the probability of propagation is determined by inputting a theme, topic, or keyword of the input message to the predictive model.

15. A method performed using at least one computer-based processor, the method comprising:

monitoring messages communicated among various computing devices over a computer network using Internet Protocol to determine whether messages are propagated among the various computing devices, the monitoring comprising collecting propagation metadata for instances where individual messages are propagated;

training a predictive model using the propagation metadata; and

predicting propagation of an input message using the predictive model.

16. The method of claim 15, further comprising:

collecting non-propagation metadata for other instances where other individual messages are not propagated; and

training the predictive model using the non-propagation metadata.

17. The method of claim 15, wherein the messages and the input message are communicated using a social network.

18. The method of claim 15, wherein the predicting comprises determining a probability that the input message will be propagated.

19. The method of claim 15, wherein the predicting comprises estimating a number of users that will receive the input message.

20. The method of claim 15, wherein the training comprises training the predictive model using content of the individual messages that are propagated and the predicting comprises inputting content of the input message to the predictive model.