Communication network optimization based on predicted enhancement gain

1. A method comprising, by one or more computing systems:

accessing data samples associated with a geographic area of interest covered by a communication network, wherein the data samples are aggregated into a plurality of data points;

partitioning the plurality of data points into a first set of data points and a second set of data points using a first threshold of a first network metric;

determining a trend of a second network metric with respect to the first network metric based on a regression analysis of the first and second sets of data points;

determining a predicted gain of the second network metric for a network enhancement operation, wherein the predicted gain of the second network metric is determined based on a difference between the trend of the second network metric and a predicted value of the second network metric after the network enhancement operation with respect to a reference value of the first network metric; and

generating one or more network optimization recommendations for the geographic area of interest based at least in part on the predicted gain of the second network metric caused by the network enhancement operation.

2. The method of claim 1 , further comprising:

determining a predicted value of the first network metric after the network enhancement operation based on a simplified cell model;

determining an intermediate predicted value of the second network metric after the network enhancement operation based on the trend of the second network metric and the predicted value of the first network metric after the network enhancement operation; and

determining the predicted value of the second network metric of the area of interest after the network enhancement operation based on the intermediate predicted value of the second network metric after the network enhancement operation.

3. The method of claim 2 , wherein the reference value of the first network metric is a measured value of the first network metric at current time, and wherein the predicted gain of the second network metric is determined by comparing the predicted value of the second network metric after the network enhancement operation to a current value of the second network metric before the network enhancement operation.

4. The method of claim 2 , further comprising:

determining a mid-term or long-term trend of the first network metric over a period of time;

determining a first predicted future value of the first network metric of a future time based on the mid-term or long-term trend of the first network metric; and

determining a second predicted future value of the second network metric of the future time based on the first predicted future value of the first network metric of the future time.

5. The method of claim 4 , wherein the reference value of the first network metric is the predicted future value of the first network metric of the future time, and wherein the predicted gain of the second network metric is determined by comparing the predicted value of the second network metric after the network enhancement operation at the future time to the second predicted future value of the second network metric before the network enhancement operation of the future time.

6. The method of claim 2 , wherein the first network metric is number of samples, and wherein the second network metric is network traffic.

7. The method of claim 6 , wherein the network enhancement operation comprises adding a new cell in the geographic area of interest, wherein the network traffic is equally split between an existing cell of the geographic area of interest and the new cell based on the simplified cell model.

8. The method of claim 1 , further comprising:

optimizing the communication network in the geographic area of interest based on the one or more network optimization recommendations.

9. The method of claim 1 , wherein the data samples are collected at application level or infrastructure level, and wherein the data samples comprise information related at least to the first and second network metrics.

10. The method of claim 1 , wherein the plurality of data points is aggregated per hour per N days, and wherein N is any positive integer number.

11. The method of claim 10 , wherein the plurality of data points is aggregated per hour per week, and wherein the plurality of data points correlates the first network metric to the second network metric.

12. The method of claim 1 , wherein the first threshold is a 50-percentile threshold of the first network metric.

13. The method of claim 1 , wherein the first threshold is a median value threshold of the first network metric.

14. The method of claim 1 , wherein the first set of data points is below the first threshold of the first network metric and corresponds to non-busy hours, and wherein the second set of data points is above the first threshold and corresponds to busy hours.

15. The method of claim 1 , wherein the trend of the second network metric with respect to the first network metric is determined based on a first trend function and a second trend function, wherein the first trend function is determined based on a first regression on the first set of data points, wherein the second trend function is determined based on a second regression on the second set of data points, and wherein the first and second regression are linear regression or non-linear regression.

16. The method of claim 1 , wherein the first network metric is number of samples, and wherein the second network metric is download speed.

17. The method of claim 16 , further comprising:

splitting the plurality of aggregated data points into the first set of data points and the second set of data points using a second threshold of the first network metric, wherein the first set of data point is below the second threshold of the first network metric;

determining a first weighted average of the second network metric based on the plurality of data points comprising the first and second sets of data points; and

determining a second weighted average of the second network metric based on the first set of data points, wherein the predicted gain of the second network metric is determined by comparing the first and second weighted average of the second network metric.

18. The method of claim 17 , wherein the second threshold of the first network metric equal to 0.5 times of a maximum value of the first network metric.

19. One or more computer-readable non-transitory storage media embodying software that is operable when executed by one or more processors to:

access data samples associated with an area of interest covered by a communication network, wherein the data samples are aggregated into a plurality of data points;

partition the plurality of data points into a first set of data points and a second set of data points using a first threshold of a first network metric;

determine a trend of a second network metric with respect to the first network metric based on a regression analysis on the first and second sets of data points;

determine a predicted gain of the second network metric for a network enhancement operation, wherein the predicted gain of the second network metric is determined based on a difference between the trend of the second network metric and a predicted value of the second network metric after the network enhancement operation with respect to a reference value of the first network metric; and

generate one or more network optimization recommendations for the geographic area of interest based at least in part on the predicted gain of the second network metric caused by the network enhancement operation.

20. A system comprising: one or more non-transitory computer-readable storage media embodying instructions; and one or more processors coupled to the storage media and operable to execute the instructions to:

access data samples associated with a geographic area of interest covered by a communication network, wherein the data samples are aggregated into a plurality of data points;

partition the plurality of data points into a first set of data points and a second set of data points using a first threshold of a first network metric;

determine a trend of a second network metric with respect to the first network metric based on a regression analysis on the first and second sets of data points;

determine a predicted gain of the second network metric for a network enhancement operation, wherein the predicted gain of the second network metric is determined based on a difference between the trend of the second network metric and a predicted value of the second network metric after the network enhancement operation with respect to a reference value of the first network metric; and

generate one or more network optimization recommendations for the geographic area of interest based at least in part on the predicted gain of the second network metric caused by the network enhancement operation.