Machine-learning framework for spectrum allocation
The disclosed embodiments provide improved methods and systems for predicting spectrum allocation in a wireless access network. The method may comprise receiving data associated with an access point (AP), creating, based on the received data and a machine learning algorithm, a decision model for predicting whether to allocate spectrum to the AP, generating a prediction, using the decision model, whether to allocate spectrum to the AP, and sending the prediction to a spectrum access system (SAS) in the wireless access network. The SAS may be configured to allocate spectrum to the AP based on the prediction.
1. A method for predicting spectrum allocation in a wireless access network, the method comprising:
receiving data associated with an access point (AP), wherein the data comprises social media inputs, and wherein the social media inputs include information reported by a user in connection with one or more social media accessed by the user through the AP;
creating, based on the received data and a machine learning algorithm, a decision model for predicting whether to allocate spectrum to the AP;
generating a prediction, using the decision model, whether to allocate spectrum to the AP; and
sending the prediction to a spectrum access system (SAS) in the wireless access network, wherein the SAS is configured to allocate spectrum to the AP based on the prediction.
2. The method of claim 1 , wherein the machine learning algorithm comprises a classification model.
3. The method of claim 1 , wherein the decision model for predicting whether to allocate spectrum to the AP comprises at least one of Decision Tree, C4. 5 algorithm, Classification and Regression Trees (CART), or Random Forest.
4. The method of claim 1 , wherein the received data associated with the AP further comprises at least one of SAS analytics, network analytics, or user device and application analytics.
5. The method of claim 4 , wherein the SAS analytics comprise at least one of spectrum request history, nature of spectrum allocation, or spectrum denial history.
6. The method of claim 4 , wherein the networks analytics comprise at least one of operation history associated with the AP, load history associated with the AP, network details associated with the AP, performance of the AP, or connection history of the AP.
7. The method of claim 4 , wherein the user device and application analytics comprise at least one of application performance, user equipment (UE) statistics, or battery life statistics.
8. The method of claim 1 , wherein the social media inputs comprise at least one of location-based information, time-based information, emergency- or disaster-related information, or information associated with poor user experience reported by users.
9. The method of claim 1 , further comprising:
monitoring key performance indicators (KPIs) associated with the AP; and
using the machine learning algorithm to update the decision model based on the monitored KPIs.
10. The method of claim 1 , further comprising:
receiving a request for more spectrum from the AP;
using the decision model to determine whether the SAS should allocate more spectrum to the AP; and
sending a notification to the SAS in the wireless access network with a recommendation regarding whether the SAS should allocate more spectrum to the AP.
11. The method of claim 10 , wherein the received data associated with the AP further comprises a ratio of average-system-throughput after a first spectrum allocation to average-system-throughput before the first spectrum allocation, and a ratio of average-system-throughput after a second spectrum allocation to average-system-throughput before the second spectrum allocation.
12. The method of claim 11 , wherein the received data associated with the AP further comprises back-haul congestion data and at least one of SAS analytics, network analytics, or user device and application analytics.
13. The method of claim 10 , further comprising:
monitoring key performance indicators (KPIs) associated with the AP; and
using the machine learning algorithm to update the decision model based on the monitored KPIs.
14. The method of claim 1 , further comprising:
receiving an indication of at least one of an incumbent interference event, a coverage outage, or a capacity outage associated with the AP;
analyzing, based on an attribute selection algorithm, the received data associated with the AP;
determining an attribute, from the received data, that is most likely to have caused at least one of the incumbent interference event, the coverage outage, or the capacity outage; and
sending the attribute to an SAS optimization algorithm, wherein the SAS is configured to use the SAS optimization algorithm to optimize the attribute.
15. The method of claim 14 , wherein the attribute comprises at least one of a propagation modeling parameter, a transmission (TX) power, an interference threshold, a portion of a spectrum on which the AP is operating, an AP vendor, a network operator vendor, a network operation parameter, an antenna parameter, a bandwidth allocation, a technology capability of the AP, a floor plan type, a building type, a type of deployment scenario, or an AP location.
16. A system for predicting spectrum allocation in a wireless access network, the system comprising:
a memory storing instructions; and
a processor configured to execute the instructions to:
receive data associated with an access point (AP), wherein the data comprises social media inputs, and wherein the social media inputs include information reported by a user in connection with one or more social media accessed by the user through the AP;
create, based on the received data and a machine learning algorithm, a decision model for predicting whether to allocate spectrum to the AP;
generate a prediction, using the decision model, whether to allocate spectrum to the AP; and
send the prediction to a spectrum access system (SAS) in the wireless access network, wherein the SAS is configured to allocate spectrum to the AP based on the prediction.
17. The system of claim 16 , wherein the machine learning algorithm comprises a classification model.
18. The system of claim 16 , wherein the decision model for predicting whether to allocate spectrum to the AP comprises at least one of Decision Tree, C4. 5 algorithm, Classification and Regression Trees (CART), or Random Forest.
19. The system of claim 16 , wherein the received data associated with the AP further comprises at least one of SAS analytics, network analytics, or user device and application analytics.
20. The system of claim 19 , wherein the SAS analytics comprise at least one of spectrum request history, nature of spectrum allocation, or spectrum denial history.
21. The system of claim 19 , wherein the networks analytics comprise at least one of operation history associated with the AP, load history associated with the AP, network details associated with the AP, performance of the AP, or connection history of the AP.
22. The system of claim 19 , wherein the user device and application analytics comprise at least one of application performance, user equipment (UE) statistics, or battery life statistics.
23. The system of claim 16 , wherein the social media inputs comprise at least one of location-based information, time-based information, emergency- or disaster-related information, or information associated with poor user experience reported by users.
24. The system of claim 16 , wherein the processor is further configured to execute the instructions to:
monitor key performance indicators (KPIs) associated with the AP; and
use the machine learning algorithm to update the decision model based on the monitored KPIs.
25. The system of claim 16 , wherein the processor is further configured to execute the instructions to:
receive a request for more spectrum from the AP;
using the decision model to determine whether the SAS should allocate more spectrum to the AP; and
send a notification to the SAS in the wireless access network with a recommendation regarding whether the SAS should allocate more spectrum to the AP.
26. The system of claim 25 , wherein the received data associated with the AP further comprises a ratio of average-system-throughput after a first spectrum allocation to average-system-throughput before the first spectrum allocation, and a ratio of average-system-throughput after a second spectrum allocation to average-system-throughput before the second spectrum allocation.
27. The system of claim 26 , wherein the received data associated with the AP further comprises back-haul congestion data and at least one of SAS analytics, network analytics, or user device and application analytics.
28. The system of claim 25 , wherein the processor is further configured to execute the instructions to:
monitor key performance indicators (KPIs) associated with the AP; and
use the machine learning algorithm to update the decision model based on the monitored KPIs.
29. The system of claim 16 , wherein the processor is further configured to execute the instructions to:
receive an indication of at least one of an incumbent interference event, a coverage outage, or a capacity outage associated with the AP;
analyze, based on an attribute selection algorithm, the received data associated with the AP;
determine an attribute, from the received data, that is most likely to have caused at least one of the incumbent interference event, the coverage outage, or the capacity outage; and
send the attribute to an SAS optimization algorithm, wherein the SAS is configured to use the SAS optimization algorithm to optimize the attribute.
30. The system of claim 29 , wherein the attribute comprises at least one of a propagation modeling parameter, a transmission (TX) power, an interference threshold, a portion of a spectrum on which the AP is operating, an AP vendor, a network operator vendor, a network operation parameter, an antenna parameter, a bandwidth allocation, a technology capability of the AP, a floor plan type, a building type, a type of deployment scenario, or an AP location.