Enabling event prediction as an on-device service for mobile interaction

Patent metadata

Field	Value
Publication number	US-9372898-B2
Application number	US-201414334631-A
Country	US
Kind code	B2
Filing date	Jul 17, 2014
Priority date	Jul 17, 2014
Publication date	Jun 21, 2016
Grant date	Jun 21, 2016

Abstract

By knowing which upcoming actions a user might perform, a mobile application can optimize a user interface or reduce the amount of user input needed for accomplishing a task. A herein-described prediction module can answer queries from a mobile application regarding which actions in the application the user is likely to perform at a given time. Any application can register and communicate with the prediction module via a straightforward application programming interface (API). The prediction module continuously learns a prediction model for each application based on the application's evolving event history. The prediction module generates predictions by combining multiple predictors with an online learning method, and capturing event patterns not only within but also across registered applications. The prediction module is evaluated using events collected from multiple types of mobile devices.

First claim

What is claimed is:

1. A method, comprising:
receiving a plurality of event notifications associated with a plurality of applications at a prediction module of a computing device, wherein the plurality of applications comprise a first application, wherein the plurality of event notifications comprise a first event notification reporting occurrence of a first event associated with the first application, and wherein the occurrence of the first event is reported based on a hierarchical naming scheme;
determining a plurality of event-prediction features for the plurality of events at the prediction module of the computing device, wherein the plurality of event-prediction features comprise a first event-prediction feature that uses a first prediction technique to determine a first value that is based on a probability of occurrence of a second event given the first event notification and a second event-prediction feature that uses a second prediction technique to determine a second value based on the probability of occurrence of the second event given the first event notification, and wherein the first prediction technique differs from the second prediction technique;
receiving a query to predict an occurrence of the second event at the prediction module of the computing device;
after receiving the query, the prediction module:
determining a sum of at least the first value and the second value, the first value determined by the first event-prediction feature using the first prediction technique, and the second value determined by the second event-prediction feature using the second prediction technique; and
determining a score predicting the occurrence of the second event based on the sum of at least the first value and the second value; and

providing the score using the prediction module of the computing device.

2. The method of claim 1 , wherein determining the sum of at least the first value and the second value comprises determining a weighted sum of at least the first value and the second value by at least determining a first weight associated with the first score based on a score related to actual occurrence of the second event.

3. The method of claim 1 , wherein the hierarchical naming scheme enables extraction of one or more sub-events from the first event.

4. The method of claim 1 , wherein the plurality of event-prediction features comprise at least one feature selected from the group of features consisting of a recency feature, a frequency feature, a Markov feature, a Poisson feature, a location feature, and a time feature.

5. The method of claim 4 , wherein the recency feature is based on a recency ranking of the first event.

6. The method of claim 4 , wherein the frequency feature is based on a frequency of occurrence of the first event.

7. The method of claim 4 , wherein the Markov feature is based on a probability of the occurrence of the first event given an occurrence of a third event, wherein the third event is an event associated with the first application.

8. The method of claim 4 , wherein the Poisson feature is based on a probability of the occurrence of the first event based on an expected occurrence frequency of the first event.

9. The method of claim 4 , wherein the location feature comprises a probability of the occurrence of the first event based on the computing device being at a predetermined location.

10. The method of claim 4 , wherein the time feature comprises a probability of the occurrence of the first event based on a predetermined time of day.

11. The method of claim 1 , further comprising:
receiving a second query to predict an occurrence of a third event at the prediction module; and
after receiving the second query, determining a score predicting the occurrence of the third event based on the plurality of event-prediction features using the prediction module.

12. The method of claim 1 , wherein a given event-prediction feature of the plurality of event-prediction features utilizes a third value that is based on a probability of an occurrence of the second event given an occurrence of the first event.

13. A computing device, comprising:
one or more processors;
a non-transitory computer readable medium configured to store executable instructions for at least a prediction module, wherein the executable instructions, when executed by the one or more processors, cause the computing device to perform functions comprising:
receiving a plurality of event notifications associated with a plurality of applications at the prediction module, wherein the plurality of applications comprise a first application, wherein the plurality of event notifications comprise a first event notification reporting occurrence of a first event associated with the first application, and wherein the occurrence of the first event is reported based on a hierarchical naming scheme,
determining a plurality of event-prediction features for the plurality of events at the prediction module, wherein the plurality of event-prediction features comprise a first event-prediction feature that uses a first prediction technique to determine a first value that is based on a probability of an occurrence of a second event given the first event notification and a second event-prediction feature that uses a second prediction technique to determine a second value based on the probability of occurrence of the second event given the first event notification, and wherein the first prediction technique differs from the second prediction technique,
receiving, at the prediction module, a query to predict an occurrence of the second event,
after receiving the query, the prediction module:
determining a sum of at least the first value and the second value, the first value determined by the first event-prediction feature using the first prediction technique, and the second value determined by the second event-prediction feature using the second prediction technique; and
determining a score predicting the occurrence of the first event based on the sum of the at least the first value and the second value, and

providing the score using the prediction module.

14. The computing device of claim 13 , wherein determining the sum of at least the first value and the second value comprises determining a weighted sum of at least the first value and the second value.

15. The computing device of claim 13 , wherein the plurality of event-prediction features comprise at least one feature selected from the group of features consisting of a recency feature, a frequency feature, a Markov feature, a Poisson feature, a location feature, and a time feature.

16. The computing device of claim 13 , further comprising:
receiving a second query to predict an occurrence of a third event at the prediction module; and
after receiving the second query, determining a score predicting the occurrence of the third event based on the plurality of event-prediction features using the prediction module.

17. The computing device of claim 13 , wherein a given event-prediction feature of the plurality of event-prediction features utilizes a third value that is based on a probability of an occurrence of the second event given an occurrence of the first event.

18. A non-transitory computer readable medium configured to store at least executable instructions, wherein the executable instructions, when executed by one or more processors of a computing device, cause the computing device to perform functions comprising:
receiving a plurality of event notifications associated with a plurality of applications at a pred

Assignees

• Google Inc

Inventors

• Li Yang

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06F9/542 Primary

  Event management; Broadcasting; Multicasting; Notifications · CPC title

• G06N5/02

  Knowledge representation; Symbolic representation · CPC title

• G06F16/24578 Primary

  using ranking · CPC title

• G06F17/3053 Primary

  Physics · mapped topic