Systems and methods for estimating user attention

1. A computer implemented method comprising:

identifying, by a computer system, first activity data for a first user of a website or an application, wherein the first activity data indicates activities of the first user on the website or the application, the website or the application associated with a window of a plurality of windows displayable on a device associated with the first user;

randomly selecting, by the computer system, a first predetermined period of inactivity from an exponential distribution, where the first predetermined period of inactivity is variable to constitute training data to train a machine learning model to predict whether users are passively present on the website or the application;

detecting, by the computer system, the first predetermined period of inactivity in the first activity data;

initiating, by the computer system, a response triggering event designed to trigger a user response from the first user after the first predetermined period of inactivity, wherein the response triggering event designed to trigger the user response is applied to the window associated with the website or the application;

monitoring, by the computer system, for an indication of the user response to the response triggering event, the user response relating to the window associated with the website or the application; and

determining, by the computer system, whether the first user is passively present on the website or the application associated with the window of the plurality of windows based on the monitoring for the indication of the user response to the response triggering event designed to trigger the user response.

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

determining, by the computer system, the first user is present on the website or the application when the indication of the user response is detected within a threshold time period for response.

3. The computer implemented method of claim 2 , wherein the first user is determined to be passively present on the website or the application at a time when the response triggering event is initiated.

4. The computer implemented method of claim 1 , further comprising: determining, by the computer system, the first user is absent from the website or the application when the indication of the user response is not detected within a threshold time period for response.

5. The computer implemented method of claim 4 , wherein the first user is determined to be absent from the website or the application at a time when the response triggering event is initiated.

6. The computer implemented method of claim 1 , further comprising: logging, by the computer system, whether the first user is determined to be passively present on the website or the application.

7. The computer implemented method of claim 6 , further comprising: logging, by the computer system, contextual information associated with the logging.

8. The computer implemented method of claim 7 , wherein the contextual information comprises one or more of the following: a page of the website or the application the first user was on, a browser on which the website or the application was run, and an operating system of a device on which the website or the application was run.

9. The computer implemented method of claim 8 , wherein the contextual information comprises one or more of the following: a type of client device, a time of day or week, and demographic information.

10. The computer implemented method of claim 1 , wherein the response triggering event is a dimming of the website or the application.

11. The computer implemented method of claim 1 , wherein the response triggering event is a displaying of a request for the first user to confirm whether the user is present on the website or the application.

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

training, by the computer system, the model to predict whether users are passively present on the website or the application based at least in part on the determining whether the first user is passively present.

13. The computer implemented method of claim 12 , wherein the model is a regression model.

14. The computer implemented method of claim 12 , wherein the model is a threshold model.

15. The computer implemented method of claim 12 , further comprising:

identifying, by the computer system, second activity data for a second user, wherein the second activity data indicates activities of the second user on the website or the application;

detecting, by the computer system, a first interval of inactivity in the second activity data;

generating, by the computer system, a prediction of whether the second user is passively present on the website or the application during the first interval of inactivity based on the model; and

estimating, by the computer system, a time the second user is passively present on the website or the application during the first interval of inactivity based on the prediction.

16. The computer implemented method of claim 15 , wherein the prediction is a probability.

17. The computer implemented method of claim 15 , wherein a total time the second user is determined to be on the website or the application comprises: the time the second user is passively present on the website or the application during the first interval of inactivity; and a time the second user is determined to be actively present on the website or the application.

18. The computer implemented method of claim 1 , wherein the computer system includes a server associated with a social networking system.

19. A system comprising:

at least one processor; and

a memory storing instructions configured to instruct the at least one processor to perform:

identifying first activity data for a first user of a website or an application, wherein the first activity data indicates activities of the first user on the website or the application, the website or the application associated with a window of a plurality of windows displayable on a device associated with the first user;

randomly selecting a first predetermined period of inactivity from an exponential distribution, where the first predetermined period of inactivity is variable to constitute training data to train a machine learning model to predict whether users are passively present on the website or the application;

detecting the first predetermined period of inactivity in the first activity data based on the bit array;

initiating a response triggering event designed to trigger a user response from the first user after the first predetermined period of inactivity, wherein the response triggering event designed to trigger the user response is applied to the window associated with the website or the application;

monitoring for an indication of the user response to the response triggering event, the user response relating to the window associated with the website or the application; and

determining whether the first user is passively present on the website or the application associated with the window of the plurality of windows based on the monitoring for the indication of the user response to the response triggering event designed to trigger the user response.

20. Anon-transitory computer storage medium storing computer-executable instructions that, when executed, cause a computer system to perform a computer-implemented method comprising:

identifying first activity data for a first user of a website or an application, wherein the first activity data indicates activities of the first user on the website or the application, the website or the application associated with a window of a plurality of windows displayable on a device associated with the first user;

randomly selecting a first predetermined period of inactivity from an exponential distribution, where the first predetermined period of inactivity is variable to constitute training data to train a machine learning model to predict whether users are passively present on the website or the application;

detecting the first predetermined period of inactivity in the first activity data based on the bit array;

initiating a response triggering event designed to trigger a user response from the first user after the first predetermined period of inactivity, wherein the response triggering event designed to trigger the user response is applied to the window associated with the website or the application;

monitoring for an indication of the user response to the response triggering event, the user response relating to the window associated with the website or the application; and

determining whether the first user is passively present on the website or the application associated with the window of the plurality of windows based on the monitoring for the indication of the user response to the response triggering event designed to trigger the user response.