Anomaly detection in a network based on a key performance indicator prediction model

1. A method, comprising:

obtaining, by a device, a measurement of a particular value of a key performance indicator (KPI) and one or more parameters of the particular value of the KPI;

determining, by the device, a prediction of the particular value of the KPI,

wherein the prediction of the particular value of the KPI comprises an output of a KPI prediction model based on providing, by the device, the one or more parameters of the particular value of the KPI as inputs to the KPI prediction model,

wherein the KPI prediction model is trained based on measurements of historical values of the KPI and associated parameters;

determining, by the device, an amount of error in the prediction of the particular value of the KPI,

wherein the amount of error in the prediction of the particular value of the KPI is based on a difference between the prediction of the particular value of the KPI and the measurement of the particular value of the KPI; and

performing, by the device and based on the amount of error in the prediction of the particular value of the KPI, one or more actions,

wherein performing the one or more actions include:

determining a weighting for a weighted average of the amount of error in the prediction of the particular value of the KPI and amounts of respective errors in predictions of the historical values of the KPI, and

updating, based on determining the weighting, a prediction accuracy value based on the amount of error in the prediction of the particular value of the KPI.

2. The method of claim 1 , further comprising:

determining the prediction accuracy value,

wherein the prediction accuracy value is based on historical amounts of errors associated with the predictions of the historical values of the KPI;

determining an anomaly threshold based on the prediction accuracy value; and

performing an action, of the one or more actions, based on the amount of error in the prediction of the particular value of the KPI satisfying the anomaly threshold.

3. The method of claim 1 ,

wherein the prediction accuracy value is based on historical amounts of errors in the predictions of the historical values of the KPI.

4. The method of claim 3 , wherein the one or more actions further comprise:

updating an anomaly threshold based on updating the prediction accuracy value,

wherein satisfaction of the anomaly threshold indicates that the particular value of the KPI is anomalous.

5. The method of claim 3 ,

wherein the weighting is based on respective ages of the predictions of the historical values of the KPI.

6. The method of claim 1 , further comprising:

determining that the amount of error in the prediction of the particular value of the KPI fails to satisfy an anomaly threshold,

wherein, based on the amount of error in the prediction of the particular value of the KPI failing to satisfy the anomaly threshold, the one or more actions comprise providing the measurement of the particular value of the KPI and the one or more parameters for further training of the KPI prediction model.

7. The method of claim 1 , further comprising:

determining that the amount of error in the prediction of the particular value of the KPI satisfies an anomaly threshold,

wherein the one or more actions comprise providing, to a network operator device and based on the amount of error in the prediction of the particular value of the KPI satisfying the anomaly threshold, an indication that the particular value of the KPI is anomalous.

8. The method of claim 1 , further comprising:

determining that the amount of error in the prediction of the particular value of the KPI satisfies a first anomaly threshold; and

determining that the amount of error in the prediction of the particular value of the KPI satisfies a second anomaly threshold,

wherein satisfaction of the second anomaly threshold indicates that the particular value of the KPI is relatively severely anomalous, and

wherein the one or more actions comprise providing, to a network operator device and based on the amount of error in the prediction of the particular value of the KPI satisfying the second anomaly threshold, an indication that the particular value of the KPI is relatively severely anomalous.

9. A device, comprising:

one or more memories; and

one or more processors communicatively coupled to the one or more memories, configured to:

train a key performance indicator (KPI) prediction model, based on measurements of historical values of a KPI and associated parameters, using one or more machine learning processes;

obtain a measurement of a particular value of the KPI and one or more parameters of the particular value of the KPI;

determine, using the KPI prediction model, a prediction of the particular value of the KPI;

determine an amount of error in the prediction of the particular value of the KPI,

wherein the amount of error in the prediction of the particular value of the KPI is based on a difference between the prediction of the particular value of the KPI and the measurement of the particular value of the KPI; and

perform, based on the amount of error in the prediction of the particular value of the KPI, one or more actions,.

wherein the one or more processors, when performing the one or more actions, are configured to:

determine a weighting for a weighted average of the amount of error in the prediction of the particular value of the KPI and amounts of respective errors in predictions of the historical values of the KPI, and

update, based on determining the weighting, a prediction accuracy value based on the amount of error in the prediction of the particular value of the KPI.

10. The device of claim 9 , wherein the one or more processors are further configured to:

determine that the prediction accuracy value satisfies an accuracy threshold,

wherein the prediction accuracy value is based on amounts of errors associated with the predictions of the historical values of the KPI; and

determine, based on the prediction accuracy value satisfying the accuracy threshold, to use the KPI prediction model to determine the prediction of the particular value of the KPI.

11. The device of claim 9 , wherein the prediction accuracy value is based on historical amounts of errors associated with the predictions of the historical values of the KPI; and

wherein the one or more processors are further configured to:

determine an anomaly threshold based on the prediction accuracy value; and

perform an action of the one or more actions based on the amount of error satisfying the anomaly threshold.

12. The device of claim 9 ,

wherein the prediction accuracy value is based on historical amounts of errors in the predictions of the historical values of the KPI.

13. The device of claim 12 , wherein the one or more actions comprise:

updating an anomaly threshold based on updating the prediction accuracy value,

wherein satisfaction of the anomaly threshold indicates that the particular value of the KPI is anomalous.

14. The device of claim 9 , wherein the one or more processors are further configured to:

determine that the amount of error in the prediction of the particular value of the KPI fails to satisfy an anomaly threshold,

wherein, based on the amount of error in the prediction of the particular value of the KPI failing to satisfy the anomaly threshold, the one or more actions comprise further training the KPI prediction model based on the measurement of the particular value of the KPI and the one or more parameters.

15. The device of claim 9 , wherein the one or more processors are further configured to:

determine that the amount of error in the prediction of the particular value of the KPI satisfies an anomaly threshold,

wherein the one or more actions comprise providing, to a network operator device and based on the amount of error in the prediction of the particular value of the KPI satisfying the anomaly threshold, an indication that the particular value of the KPI is anomalous.

16. The device of claim 9 , wherein the one or more parameters of the particular value of the KPI comprise one or more of:

a time of day associated with the measurement of the particular value of the KPI;

a day of a week associated with the measurement of the particular value of the KPI;

a date associated with the measurement of the particular value of the KPI;

whether the measurement of the particular value of the KPI occurred on a holiday; or

whether the measurement of the particular value of the KPI occurred on a weekend.

17. A non-transitory computer-readable medium storing instructions, the instructions comprising:

one or more instructions that, when executed by one or more processors, cause the one or more processors to:

determine to use a key performance indicator (KPI) prediction model to determine a prediction of a particular value of a KPI based on a prediction accuracy value of the KPI prediction model that satisfies an accuracy threshold;

wherein the prediction accuracy value is based on amounts of errors associated with predictions of historical values of the KPI;

obtain a measurement of the particular value of the KPI and one or more parameters of the particular value of the KPI;

determine, using the KPI prediction model, the prediction of the particular value of the KPI;

determine an amount of error in the prediction of the particular value of the KPI,

wherein the amount of error is based on a difference between the prediction of the particular value of the KPI and the measurement of the particular value of the KPI; and

perform, based on the amount of error in the prediction of the particular value of the KPI, one or more actions,

wherein the one or more instructions, that cause the one or more processors to perform the one or more actions, cause the one or more processors to:

determine a weighting for a weighted average of the amount of error in the prediction of the particular value of the KPI and amounts of respective errors in the predictions of the historical values of the KPI, and

update, based on determining the weighting, the prediction accuracy value based on the amount of error in the prediction of the particular value of the KPI.

18. The non-transitory computer-readable medium of claim 17 , wherein the one or more instructions, when executed by the one or more processors, further cause the one or more processors to:

determine an anomaly threshold based on the prediction accuracy value; and

perform an action of the one or more actions based on the amount of error satisfying the anomaly threshold.

19. The non-transitory computer-readable medium of claim 17 , wherein the one or more instructions, when executed by the one or more processors, further cause the one or more processors to:

determine that the amount of error in the prediction of the particular value of the KPI fails to satisfy an anomaly threshold,

wherein the one or more instructions, that cause the one or more processors to perform the one or more actions, cause the one or more processors to provide, based on the amount of error in the prediction of the particular value of the KPI failing to satisfy the anomaly threshold, the measurement of the particular value of the KPI and the one or more parameters for further training for the KPI prediction model.

20. The non-transitory computer-readable medium of claim 17 , wherein the one or more instructions, when executed by the one or more processors, further cause the one or more processors to:

determine that the amount of error in the prediction of the particular value of the KPI satisfies an anomaly threshold,

wherein the one or more instructions, that cause the one or more processors to perform the one or more actions, cause the one or more processors to provide, to a network operator device and based on the amount of error in the prediction of the particular value of the KPI satisfying the anomaly threshold, an indication that the particular value of the KPI is anomalous.