Automatically optimize parameters via machine learning

1. A computer-implemented method for configuring parameters in a wireless communications network having one or more cells, the method comprising:

determining whether parameter configurations result in a change to at least one of key quality indicator (KQI) or key performance indicator (KPI) measurements based on data samples collected in the wireless communications network;

dividing the data samples into subsets including

a first subset of the data samples including the data samples associated with the parameter configurations failing to result in the change to the at least one of the KQI or KPI measurements, and

a second subset of the data samples including the data samples associated with the parameter configurations resulting in the change to at least one of the KQI or KPI measurements dependent upon a triggering threshold for an operation being satisfied and the operation satisfying the triggering threshold being successfully executed in the wireless communications network;

determining which of the subsets of the data samples to use for optimizing the parameter configurations; and

providing the determined subsets of the data samples as an input to machine learning to learn the parameter configurations that yield the at least one of the KQI or KPI measurements to optimize performance in the wireless communications network, the providing comprising

selecting features for clustering the one or more cells in the wireless communications network as a predictor of performance such that the selected features result in a change to the at least one of the KQI or KPI measurements;

identifying the features that are independent of the parameter configurations to be optimized;

clustering the one or more cells for a time period using the identified features; and

selecting, within a cluster of the one or more cells, values for the parameter configurations that optimize the performance in the wireless communications network.

2. The method of claim 1 , wherein the determination of whether the parameter configurations result in a change to at least one of the KQI or KPI measurements is based on satisfying one or more conditions in the wireless communications network.

3. The method of claim 2 , wherein the one or more conditions include at least of a triggering threshold, an operation being triggered and the operation being successfully executed.

4. The method of claim 1 , further comprising:

separating the parameter configurations into categories including

a first category corresponding to the data samples for which the parameter configurations always result in the change to at least one of the KQI or KPI measurements, and

a second category corresponding to the first and second subsets of the data samples; and

providing the first category of parameter configurations as an input to machine learn the parameter configurations to optimize the wireless communications network.

5. The method of claim 1 , wherein the parameter configurations include at least one of a single parameter, a composite parameter including multiple parameters, a function of one or multiple parameters, and an expression of one or multiple parameters.

6. The method of claim 1 , wherein the first and second subsets of the data samples learn from each other.

7. The method of claim 1 ,

wherein selecting the values comprises:

comparing the wireless communications network performance for each of the values of parameter configurations, and

choosing the value associated with the parameter configuration that results in the optimized wireless communications network performance; and

further comprising:

reconfiguring the one or more cells in the cluster in which the wireless communication network results in less than optimal network performance using the chosen value.

8. The method of claim 1 , wherein identifying the features as independent of the parameter configurations to be optimized is determined by at least one of the effect of changing the parameter configurations of the feature, a degree of impact of the feature changing when the parameter configurations are tuned, or a correlation of the parameter configurations and the feature.

9. A device for configuring parameters in a wireless communications network having one or more cells, comprising:

a non-transitory memory storage comprising instructions; and

one or more processors in communication with the memory, wherein the one or more processors execute the instructions to perform operations comprising:

determining whether parameter configurations result in a change to at least one of key quality indicator (KQI) or key performance indicator (KPI) measurements based on data samples collected in the wireless communications network;

dividing the data samples into subsets including

a first subset of the data samples including the data samples associated with the parameter configurations failing to result in the change to the at least one of the KQI or KPI measurements, and

a second subset of the data samples including the data samples associated with the parameter configurations resulting in the change to at least one of the KQI or KPI measurements dependent upon a triggering threshold for an operation being satisfied and the operation satisfying the triggering threshold being successfully executed in the wireless communications network;

determining which of the subsets of the data samples to use for optimizing the parameter configurations; and

providing the determined subsets of the data samples as an input to machine learning to learn the parameter configurations that yield the at least one of the KQI or KPI measurements to optimize performance in the wireless communications network, the providing comprising

selecting features for clustering the one or more cells in the wireless communications network as a predictor of performance of the wireless communication network such that the selected features result in a change to the at least one of the KQI or KPI measurements;

identifying the features that are independent of the parameter configurations to be optimized;

clustering the one or more cells for a time period using the identified features; and

selecting, within a cluster of the one or more cells, values for the parameter configurations that optimize the performance in the wireless communications network.

10. The device of claim 9 , wherein the determination of whether the parameter configurations result in a change to at least one of the KQI or KPI measurements is based on satisfying one or more conditions in the wireless communications network.

11. The device of claim 10 , wherein the one or more conditions include at least of a triggering threshold, an operation being triggered and the operation being successfully executed.

12. The device of claim 9 , wherein the one or more processors further execute the instructions to perform operations comprising:

separating the parameter configurations into categories including

a first category corresponding to the data samples for which the parameter configurations always result in the change to at least one of the KQI or KPI measurements, and

a second category corresponding to the first and second subsets of the data samples; and

providing the first category of parameter configurations as an input to machine learning for the parameter configurations to optimize the wireless communications network.

13. The device of claim 9 , wherein the parameter configurations include at least one of a single parameter, a composite parameter including multiple parameters, a function of one or multiple parameters, or an expression of one or multiple parameters.

14. The device of claim 9 , wherein the first and second subsets of the data samples learn from each other.

15. The device of claim 9 ,

wherein selecting the values comprises

comparing the wireless communications network performance for each of the values of parameter configurations, and

choosing the value associated with the parameter configuration that results in the optimized wireless communications network performance; and

the one or more processors execute the instructions to perform operations comprising:

reconfiguring the one or more cells in the cluster in which the wireless communication network results in less than optimal network performance using the chosen value.

16. The device of claim 9 , wherein identifying the features as independent of the parameter configurations to be optimized is determined by at least one of the effect of changing the parameter configurations of the feature, a degree of impact of the feature changing when the parameter configurations are tuned, or a correlation of the parameter configurations and the feature.

17. A non-transitory computer-readable medium storing computer instructions for configuring parameters in a wireless communications network having one or more cells, that when executed by one or more processors, perform the steps of:

determining whether parameter configurations result in a change to at least one of key quality indicator (KQI) or key performance indicator (KPI) measurements based on data samples collected in the wireless communications network;

dividing the data samples into subsets including

a first subset of the data samples including the data samples associated with the parameter configurations failing to result in the change to the at least one of the KQI or KPI measurements, and

a second subset of the data samples including the data samples associated with the parameter configurations resulting in the change to at least one of the KQI or KPI measurements dependent upon a triggering threshold for an operation being satisfied and the operation satisfying the triggering threshold being successfully executed in the wireless communications network;

determining which of the subsets of the data samples to use for optimizing the parameter configurations; and

providing the determined subsets of the data samples as an input to machine learning to learn the parameter configurations that yield the at least one of the KQI or KPI measurements to optimize performance in the wireless communications network, the providing comprising

selecting features for clustering the one or more cells in the wireless communications network as a predictor of performance of the wireless communication network such that the selected features result in a change to the at least one of the KQI or KPI measurements;

identifying the features that are independent of the parameter configurations to be optimized;

clustering the one or more cells for a time period using the identified features; and

selecting, within a cluster of the one or more cells, values for the parameter configurations that optimize the performance in the wireless communications network.

18. The non-transitory computer-readable medium of claim 17 , wherein the determination of whether the parameter configurations result in a change to at least one of the KQI or KPI measurements is based on satisfying one or more conditions in the wireless communications network.

19. The non-transitory computer-readable medium of claim 18 , wherein the one or more conditions include at least of a triggering threshold, an operation being triggered and the operation being successfully executed.

20. The non-transitory computer-readable medium of claim 17 , when further executed by one or more processors, perform the steps of:

separating the parameter configurations into categories including

a first category corresponding to the data samples for which the parameter configurations always result in the change to at least one of the KQI or KPI measurements, and

a second category corresponding to the first and second subsets of the data samples; and

providing the first category of parameter configurations as an input to machine learning for the parameter configurations to optimize the wireless communications network.

21. The non-transitory computer-readable medium of claim 17 , wherein the parameter configurations include at least one of a single parameter, a composite parameter including multiple parameters, a function of one or multiple parameters, or an expression of one or multiple parameters.

22. The non-transitory computer-readable medium of claim 17 , wherein the first and second subsets of the data samples learn from each other.

23. The non-transitory computer-readable medium of claim 17 ,

wherein selecting the values comprises:

comparing the wireless communications network performance for each of the values of parameter configurations, and

choosing the value associated with the parameter configuration that results in the optimized wireless communications network performance; and

when executed by one or more processors, perform the step of:

reconfiguring the one or more cells in the cluster in which the wireless communication network results in less than optimal network performance using the chosen value.

24. The non-transitory computer-readable medium of claim 17 , wherein identifying the features as independent of the parameter configurations to be optimized is determined by at least one of the effect of changing the parameter configurations of the feature, a degree of impact of the feature changing when the parameter configurations are tuned, or a correlation of the parameter configurations and the feature.