Methods and systems for predictive alerting

• US 9,373,078 B1
• Filed: 04/23/2012
• Issued: 06/21/2016
• Est. Priority Date: 04/21/2011
• Status: Expired due to Fees

• Current Assignee: Anametrix, Inc. (Cheq Technologies Ltd.)
• Original Assignee: Anametrix, Inc. (Cheq Technologies Ltd.)
• Inventors: Jugai, Roman, Jonsson, Andreas, Olsson, Anders
• Primary Examiner(s): Starks, Wilbert L
• Application Number:
  US13/453,926
• Time in Patent Office: 1,520 Days
• Field of Search: 706/46, 706/45
• US Class Current: 1/1
• CPC Class Codes: G06F 16/212 with details for data modelling support
• CPC Class Codes: G06F 16/24565 Triggers; Constraints
• CPC Class Codes: G06F 16/2462 Approximate or statistical queries
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: G06N 5/048 Fuzzy inferencing
• CPC Class Codes: G06Q 30/0242 Determining effectiveness of advertisements

Abstract

Systems and methods to provide predictive alerting are disclosed. Trends and characteristics in the data may be identified in advance of actual occurrence, with an expected event being predicted or extrapolated based on data and a predictive model associated with stored and/or incoming data. This may be done by, for example, developing a set of rules associated with the model that may then be used to trigger predictive alerts based on predicted trends. Various data and associated characteristics may be determined based on past data and/or analogous data taken from similarly modeled systems. For example, seasonal trends may be factored into the model so that a prediction is based on both a received set of data and expected seasonal (or other temporal) trends. Various other characteristics may also be used in the predictive model.

Claims

1. A method of providing predictive alerting, comprising:

receiving, by a computing device, web analytics data associated with one or more variables related to a web site, wherein the one or more variables related to the web site are indicative of one or more of results of an advertising campaign, a number of hits, and inventory information;

analyzing, by a computing device, the web analytics data to identify a seasonality pattern associated with the one or more variables related to the web site;

determining, using an autocorrelation function, a cyclical characteristic of the seasonality pattern;

determining one or more predictive models associated with the one or more variables related to the web site, the determining being based on the web analytics data associated with the one or more variables related to the web site;

receiving additional data associated with the one or more variables related to the web site; and

providing, based on the additional data and the one or more predictive models, an alert indicative of a predicted future state of the one or more variables related to the web site.

2. The method of claim 1, wherein the providing includes sending the alert as an email message.

3. The method of claim 1, wherein the providing includes providing the alert as a text message or a voice message.

4. The method of claim 1, wherein the providing includes posting the alert on a website.

5. The method of claim 1, wherein the one or more predictive models are based on a temporal variation in the one or more variables.

6. The method of claim 5, wherein the temporal variation is a seasonal variation.

7. The method of claim 1, wherein the one or more predictive models are based on an event associated with the one or more variables.

8. The method of claim 1, comprising:

receiving user-supplied configuration parameters associated with the alert.

9. The method of claim 8, wherein the user-supplied configuration parameters relate to a temporal configuration associated with the alert.

10. The method of claim 8, wherein the user-supplied configuration parameters relate to a prediction configuration associated with the alert.

11. The method of claim 8, wherein the user-supplied configuration parameters relate to a confidence configuration associated with the alert.

12. The method of claim 1, comprising:

determining that the web analytics data exhibits randomness,

wherein the determining the one or more predictive models is based on the web analytics data exhibiting randomness.

13. The method of claim 1, comprising:

determining, using the autocorrelation function, a number of lags in the web analytics data.

14. The method of claim 1, wherein the cyclical characteristic of the seasonality pattern comprises one of an average number of times the seasonality pattern changes over a period of time and an average interval length between two different data points at which the seasonality pattern changes.

15. The method of claim 1, comprising:

determining, using the autocorrelation function, whether the web analytics data is random.

16. The method of claim 1, comprising:

determining a confidence of the predicted future state of the one or more variables related to the web site; and

determining whether the confidence of the predicted future state of the one or more variables is greater than or equal to a defined confidence level,

wherein the alert is provided based on the confidence of the predicted future state of the one or more variables being greater than or equal to the defined confidence level.

17. The method of claim 1, comprising:

determining a predicted time that the predicted future state of the one or more variables related to the web site will occur,

wherein the alert includes the predicted time.

18. The method of claim 1, comprising:

determining a non-peak window of time of use of the computing device;

receiving a selection of at least one of the variables related to the web site, wherein analyzing the web analytics data to identify the seasonality pattern is performed using the selected at least one of the variables related to the web site;

determining an average period of time before each change in trend in the cyclical characteristic of the seasonality pattern;

determining, during the non-peak window of time, based on the average period of time before each change in trend in the cyclical characteristic of the seasonality pattern, a predicted time that the predicted future state of the one or more variables related to the web site will occur; and

sending the alert indicative of the predicted future state of the one or more variables related to the web site, wherein the alert includes the predicted time.

19. Non-transitory computer-readable media storing instructions that, when executed by a processor, cause a system to:

receive web analytics data associated with one or more variables related to a web site, wherein the one or more variables related to the web site are indicative of one or more of results of an advertising campaign, a number of hits, and inventory information;

analyze the web analytics data to identify a seasonality pattern associated with the one or more variables related to the web site;

determine, using an autocorrelation function, a cyclical characteristic of the seasonality pattern;

determine one or more predictive models associated with the one or more variables related to the web site, the determining being based on the web analytics data associated with the one or more variables related to the web site;

receive additional data associated with the one or more variables related to the web site; and

provide, based on the additional data and the one or more predictive models, an alert indicative of a predicted future state of the one or more variables related to the web site.

20. The non-transitory computer-readable media of claim 19, storing instructions that, when executed by the processor, cause the system to:

send the alert as an email message.

21. The non-transitory computer-readable media of claim 19, storing instructions that, when executed by the processor, cause the system to:

provide the alert as a text message or a voice message.

22. The non-transitory computer-readable media of claim 19, storing instructions that, when executed by the processor, cause the system to:

provide the alert on a website.

23. The non-transitory computer-readable media of claim 19, wherein the one or more predictive models are based on a temporal variation in the one or more variables.

24. The non-transitory computer-readable media of claim 23, wherein the temporal variation is a seasonal variation.

25. The non-transitory computer-readable media of claim 19, wherein the one or more predictive models are based on an event associated with the one or more variables.

26. The non-transitory computer-readable media of claim 19, storing instructions that, when executed by the processor, cause the system to:

receive configuration parameters associated with the alert.

27. The non-transitory computer-readable media of claim 26, wherein the configuration parameters relate to a temporal configuration associated with the alert.

28. The non-transitory computer-readable media of claim 26, wherein the configuration parameters relate to a prediction configuration associated with the alert.

29. The non-transitory computer-readable media of claim 26, wherein the configuration parameters relate to a confidence configuration associated with the alert.

30. The non-transitory computer-readable media of claim 26, storing instructions that, when executed by the processor, cause the system to:

determine that the web analytics data exhibits randomness,

wherein determining the one or more predictive models is based on the web analytics data exhibiting randomness.

31. The non-transitory computer-readable media of claim 19, wherein the instructions, when executed by the processor, cause the system to:

determine a non-peak window of time of use of the system;

receive a selection of at least one of the variables related to the web site, wherein analyzing the web analytics data to identify the seasonality pattern is performed using the selected at least one of the variables related to the web site;

determine an average period of time before each change in trend in the cyclical characteristic of the seasonality pattern;

determine, using the at least one processor during the non-peak window of time, based on the average period of time before each change in trend in the cyclical characteristic of the seasonality pattern, a predicted time that the predicted future state of the one or more variables related to the web site will occur; and

send the alert indicative of the predicted future state of the one or more variables related to the web site, wherein the alert includes the predicted time.

32. A system, comprising:

at least one processor; and

a memory storing instructions that, when executed by the at least one processor, cause the system to;

receive web analytics data associated with one or more variables related to a web site, wherein the one or more variables related to the web site are indicative of one or more of results of an advertising campaign, a number of hits, and inventory information;

analyze the web analytics data to identify a seasonality pattern associated with the one or more variables related to the web site;

determine, using an autocorrelation function, a cyclical characteristic of the seasonality pattern;

determine one or more predictive models associated with the one or more variables related to the web site, the determining being based on the web analytic data associated with the one or more variables related to the web site;

receive additional data associated with the one or more variables related to the web site; and

generate, based on the additional data and the one or more predictive models, an alert indicative of a predicted future state of the one or more variables related to the web site.

33. The system of claim 32, wherein the instructions, when executed by the at least one processor, cause the system to:

determine a non-peak window of time of use of the system;

receive a selection of at least one of the variables related to the web site, wherein analyzing the web analytics data to identify the seasonality pattern is performed using the selected at least one of the variables related to the web site;

determine an average period of time before each change in trend in the cyclical characteristic of the seasonality pattern;

determine, using the at least one processor during the non-peak window of time, based on the average period of time before each change in trend in the cyclical characteristic of the seasonality pattern, a predicted time that the predicted future state of the one or more variables related to the web site will occur; and

send the alert indicative of the predicted future state of the one or more variables related to the web site, wherein the alert includes the predicted time.