Providing personalized alerts and anomaly summarization

Patent metadata

Field	Value
Publication number	US-12112349-B2
Application number	US-201615238208-A
Country	US
Kind code	B2
Filing date	Aug 16, 2016
Priority date	Aug 16, 2016
Publication date	Oct 8, 2024
Grant date	Oct 8, 2024

Abstract

Methods and systems are provided herein for summarizing a set of anomalies corresponding to a group of metrics of interest to a monitoring system user. Initially, a set of anomalies corresponding to a group of metrics is identified as having values that are outside of a predetermined range. A correlation value is determined for at least a portion of pairs of anomalies in the set of anomalies. For each anomaly in the set of anomalies, an informativeness value is computed that indicates how informative each anomaly in the set of anomalies is to the monitoring system user. The correlation values and the informativeness values are then used to identify at least one key anomaly and a plurality of non-key anomalies from the set of anomalies. A summary is generated of the identified at least one key anomaly to provide information to the monitoring system user about the set of anomalies for a particular time period.

First claim

What is claimed is:

1. A method for summarizing a set of anomalies corresponding to a group of metrics of interest to a monitoring system user, the method comprising:
identifying a set of anomalies corresponding to the group of metrics as having anomaly values that are outside of a predetermined range;
determining, for each anomaly in the set of anomalies:
a severity factor indicating a magnitude of deviation between the anomaly value corresponding to an anomaly and the predetermined range;
an interest value indicating the monitoring system user's interest in the group of metrics that is based on using a number of times the monitoring system user has interacted with a prior report associated with the group of metrics in a first period of time to determine a probability that the monitoring system user will request a report associated with the group of metrics in a next period of time; and
a confidence factor representing a confidence in the interest value, for the group of metrics, based on a number of samples in a set of samples, wherein each sample of the set of samples comprises the interest value determined at a particular time within an interval of time;
determining, for each anomaly in the set of anomalies, an informativeness value that indicates how informative an anomaly in the set of anomalies is to the monitoring system user, the informativeness value based on the severity factor, the interest value, and the confidence factor corresponding with the anomaly in the set of anomalies;

assigning a correlation value for at least a portion of pairs of anomalies in the set of anomalies, the correlation value indicating an ability of a first anomaly of a pair of anomalies to capture information content of a second anomaly of the pair of anomalies;
utilizing the correlation value and the informativeness values to identify at least one key anomaly and a plurality of non-key anomalies from the set of anomalies; and
generating, based on a user profile including prior report consumption patterns, manual alerts, and explicit feedback of the monitoring system user, a summary of the identified at least one key anomaly to provide information to the monitoring system user about the set of anomalies for a particular time period, wherein the at least one key anomaly is prioritized in the summary based on the user profile.

2. The method of claim 1 , wherein determining the interest value is further based on a frequency at which the monitoring system user accesses the prior report associated with the group of metrics, and amount of time since a previous report was accessed by the monitoring system user, and a duration for which previous reports were generated for the group of metrics.

3. The method of claim 1 , wherein the at least one key anomaly is an anomaly in the set of anomalies that is representative of the set of anomalies based on the informativeness values, and wherein the plurality of non-key anomalies are a plurality of anomalies in the set of anomalies that are less representative of the set of anomalies than the at least one key anomaly based on the informativeness values.

4. The method of claim 1 , further comprising identifying a set of related anomalies for each of at least one key anomaly.

5. The method of claim 4 , wherein identifying the set of related anomalies further comprises:
for each of the at least one non-key anomaly, identifying one of the at least one key anomaly based on the correlation value between the pair of anomalies; and
associating each of the at least one non-key anomaly with the one of the at least one key anomaly that has a higher correlation value.

6. The method of claim 1 , wherein each anomaly in the set of anomalies represents a change in data for a corresponding metric.

7. The method of claim 4 , wherein metrics corresponding to the at least one key anomaly and the metrics corresponding to the set of related anomalies are causally dependent on one another.

8. The method of claim 1 , further comprising computing a first iteration of a gain value computation that is based on, at least, the compound correlation values and the informativeness values.

9. The method of claim 8 , wherein the first iteration of the gain value computation identifies a first key anomaly of the at least one key anomaly.

10. The method of claim 8 , further comprising computing a second iteration of the gain value computation wherein the at least one key anomaly comprises a second key anomaly from a second iteration of the gain value computation upon a determination that the second key anomaly meets a predetermined informativeness threshold that is less than the informativeness of the first key anomaly.

11. One or more computer storage media storing computer-usable instructions, that when used by a computing device, cause the computing device to perform a method for summarizing a set of anomalies corresponding to a group of metrics of interest to a monitoring system user, the method comprising:
determining, for each anomaly in the set of anomalies:
a severity factor indicating a magnitude of deviation between an anomaly value corresponding to an anomaly and a predetermined range;
an interest value indicating the monitoring system user's interest in the group of metrics that is based on a using a number of times the monitoring system user has interacted with a prior report associated with the group of metrics in a first period of time to determine a probability that the monitoring system user will request a report associated with the group of metrics in a next period of time; and
a confidence factor representing a confidence in the interest value, for the group of metrics, based on a number of samples in a set of samples, wherein each sample of the set of samples comprises the interest value determined at a particular time within an interval of time;

computing an informativeness value for each of the plurality of anomalies, the informativeness values calculated based on, at least, the severity factor, the interest value, and the confidence factor corresponding with the anomaly in the set of anomalies;
assigning a correlation value to at least one pair of anomalies for a plurality of anomalies, the correlation value indicative of how much information a first anomaly of the at least one pair of anomalies conveys in relation to a second anomaly of the at least one pair of anomalies;
from the plurality of anomalies, identifying at least one key anomaly using the correlation value of the at least one pair of anomalies and the informativeness values of each of the plurality of anomalies in the at least one pair of anomalies;
identifying at least one related anomaly for each of the at least one key anomaly by selecting at least one non-key anomaly having a highest correlation value with the each of the at least one key anomaly; and
generating, based on a user profile including prior actions of the monitoring system user, a message based on the at least one key anomaly and the at least one related anomaly to provide a summary for the monitoring system user of changes in data for the plurality of anomalies.

12. The one or more computer storage media of claim 11 , wherein identifying the at least one key anomaly further comprises:
computing a first iteration of a gain value for each anomaly in the plurality of anomalies using the computed correlation values and the informativeness values; and
for the first iteration of the gain value, identifying a first key anomaly of the plurality of anomalies having a highest gain value.

13. The one or more computer storage media of claim 11 , further comprising detecting the plurality of anomalies, wherein the

Assignees

• Adobe Inc

Inventors

• Modani Natwar
• Burhanuddin Iftikhar Ahamath
• Hiranandani Gaurush
• Saini Shiv Kumar

Classifications

• G06Q30/0244 Primary

  Optimization · CPC title