Granted Patent B2

US 12,739,178 · App. 18/977,679 · Granted Sep 15, 2026

Inventor: Michael C. Bascao (Renton, WA)

Assignee: T-Mobile USA, Inc.

H04L41/16H04W24/02

Abstract

Claims (62)

1 . A computer-implemented method for assessing anomalous occurrences at a telecommunications service network, the computer-implemented method comprising:

accessing, from a centralized repository, a set of telecommunications data records for a plurality of user devices operating on the telecommunications service network, wherein the set of telecommunications data records comprises (a) user device information for the plurality of user devices, (b) network performance metrics, and (c) disparate types of telecommunications data records including enhanced data records (EDRs) or call data records (CDRs);

inputting at least a portion of data from the set of telecommunications data records into a machine learning model to identify one or more anomalous occurrences, wherein the machine learning model is trained to identify anomalous occurrences from patterns in user device information and network performance metrics;

appending the one or more anomalous occurrences to a first data structure storing anomalous occurrences across the telecommunications service network;

accessing, from a plurality of different network elements of the telecommunications service network, values of metrics indicative of changes in device conditions, network operability conditions, or both;

wherein the changes in device conditions and/or network operability conditions comprise seasonal changes, device changes, location changes, user call patterns, or changes in network events or customer lifestyle;

generating a second data structure by aggregating the accessed values of the metrics indicative of changes in device conditions and/or network operability conditions;

establishing a correlation between the one or more anomalous occurrences stored in the first data structure and one or more changes in a subset of device conditions and/or network operability conditions stored in the second data structure; and

transmitting, to at least one device on the telecommunications service network, instructions configured to cause reconfiguration of one or more attributes that impact the subset of device conditions and/or network operability conditions at the at least one device.

2 . The computer-implemented method of claim 1 , wherein the one or more anomalous occurrences are occurrences of network degradation, and wherein the computer-implemented method further comprises:

generating first instructions for displaying a number of degraded sites of the telecommunications service network; generating second instructions for displaying the one or more changes in device conditions and/or network operability conditions; and transmitting, to a remote device, the first instructions and the second instructions.

3 . The computer-implemented method of claim 1 , wherein accessing the metrics indicative of the changes in device conditions and/or network operability conditions comprises accessing network elements specific to a cell site of the telecommunications service network to obtain metrics indicative of changes occurring at a site-specific level.

4 . The computer-implemented method of claim 1 ,

wherein the computer-implemented method further comprises standardizing the disparate types of telecommunications data records into a predetermined format.

5 . The computer-implemented method of claim 1 , wherein the machine learning model is a first machine learning model from a registry of machine learning models, and wherein the computer-implemented method further comprises:

periodically evaluating the machine learning models of the registry of machine learning models to obtain values for a plurality of metrics; responsive to determining that values corresponding to a model of the registry of machine learning models do not meet or exceed a predetermined threshold, removing the model from the registry temporarily; and performing refinement of the machine learning model.

6 . The computer-implemented method of claim 1 , further comprising:

accessing a training set of telecommunications data records comprising values indicative of network degradation; and training the machine learning model using the training set of telecommunications data records to identify patterns in telecommunications data records corresponding to network degradation.

7 . One or more non-transitory computer-readable media containing instructions which when executed by a processor, perform a method for assessing anomalous occurrences at a telecommunications service network, the method comprising:

accessing, from a centralized repository, a set of telecommunications data records for a plurality of user devices operating on the telecommunications service network, wherein the set of telecommunications data records comprise (a) user device information for the plurality of user devices, (b) network performance metrics, and (c) disparate types of telecommunications data records including enhanced data records (EDRs) or call data records (CDRs);

inputting at least a portion of data from the set of telecommunications data records into a machine learning model to identify one or more anomalous occurrences, wherein the machine learning model is trained to identify anomalous occurrences from patterns in user device information and network performance metrics;

appending the one or more anomalous occurrences to a first data structure storing anomalous occurrences across the telecommunications service network;

accessing, from a plurality of different network elements of the telecommunications service network, values of metrics indicative of changes in device conditions, network operability conditions, or both;

wherein the changes in device conditions and/or network operability conditions comprise seasonal changes, device changes, location changes, user call patterns, or changes in network events or customer lifestyle;

generating a second data structure by aggregating the accessed values of the metrics indicative of changes in device conditions and/or network operability conditions;

establishing a correlation between the one or more anomalous occurrences stored in the first data structure and one or more changes in a subset of device conditions and/or network operability conditions stored in the second data structure; and

transmitting, to at least one device on the telecommunications service network, instructions configured to cause reconfiguration of one or more attributes that impact the subset of device conditions and/or network operability conditions at the at least one device.

8 . The one or more non-transitory computer-readable media of claim 7 , wherein the method further comprises:

generating first instructions for displaying a number of degraded sites of the telecommunications service network; generating second instructions for displaying changes in device and network operability conditions; and transmitting, to a remote device, the first instructions and the second instructions.

9 . The one or more non-transitory computer-readable media of claim 7 , wherein accessing the metrics indicative of changes in device operability conditions comprises accessing network elements specific to a cell site of the telecommunications service network to obtain metrics indicative of changes occurring at a site-specific level.

10 . The one or more non-transitory computer-readable media of claim 7 , wherein the method further comprises standardizing the disparate types of telecommunications data records into a predetermined format.

11 . The one or more non-transitory computer-readable media of claim 7 , wherein the machine learning model is a first machine learning model from a registry of machine learning models, and wherein the method further comprises:

periodically evaluating the machine learning models of the registry of machine learning models to obtain values for a plurality of metrics; responsive to determining that values corresponding to a model of the registry of machine learning models do not meet or exceed a predetermined threshold, removing the model from the registry temporarily; and performing refinement of the machine learning model.

12 . The one or more non-transitory computer-readable media of claim 7 , wherein the method further comprises:

accessing a training set of telecommunications data records comprising values indicative of network degradation; and

training the machine learning model using the training set of telecommunications data records to identify patterns in telecommunications data records corresponding to network degradation.

13 . A system for assessing anomalous occurrences at a telecommunications service network, the system comprising:

one or more processors; and one or more non-transitory computer-readable media storing instructions that, when executed by the one or more processors, cause operations comprising:

accessing, from a centralized repository, a set of telecommunications data records for a plurality of user devices operating on the telecommunications service network, wherein the set of telecommunications data records comprise (a) user device information for the plurality of user devices (b) network performance metrics, and (c) disparate types of telecommunications data records including enhanced data records (EDRs) or call data records (CDRs);

inputting at least a portion of data from the set of telecommunications data records into a machine learning model to identify one or more anomalous occurrences, wherein the machine learning model is trained to identify anomalous occurrences from patterns in user device information and network performance metrics;

appending the one or more anomalous occurrences to a first data structure storing anomalous occurrences across the telecommunications service network;

accessing, from a plurality of different network elements of the telecommunications service network, values of metrics indicative of changes in device conditions, network operability conditions, or both;

wherein the changes in device conditions and/or network operability conditions comprise seasonal changes, device changes, location changes, user call patterns, or changes in network events or customer lifestyle;

generating a second data structure by aggregating the accessed values of the metrics indicative of changes in device conditions and/or network operability conditions;

establishing an association between the one or more anomalous occurrences stored in the first data structure and one or more changes in a subset of device conditions and/or network operability conditions stored in the second data structure; and

transmitting, to at least one device on the telecommunications service network, instructions configured to cause reconfiguration of one or more attributes that impact the subset of device conditions and/or network operability conditions at the at least one device.

14 . The system of claim 13 , wherein the one or more anomalous occurrences are occurrences of network degradation, and wherein the one or more non-transitory computer-readable media further cause operations comprising:

generating first instructions for displaying a number of degraded sites of the telecommunications service network; generating second instructions for displaying changes in device and network operability conditions; and transmitting, to a remote device, the first instructions and the second instructions.

15 . The system of claim 13 , wherein accessing the metrics indicative of changes in device operability conditions comprises accessing network elements specific to a cell site of the telecommunications service network to obtain metrics indicative of changes occurring at a site-specific level.

16 . The system of claim 13 , wherein the one or more non-transitory computer-readable media further cause operations comprising standardizing the disparate types of telecommunications data records into a predetermined format.

17 . The system of claim 13 , wherein the machine learning model is a first machine learning model from a registry of machine learning models, and wherein the one or more non-transitory computer-readable media further cause operations comprising:

periodically evaluating the machine learning models of the registry of machine learning models to obtain values for a plurality of metrics;

responsive to determining that values corresponding to a model of the registry of machine learning models do not meet or exceed a predetermined threshold,

removing the model from the registry temporarily; and performing refinement of the machine learning model.

18 . The system of claim 13 , wherein the one or more non-transitory computer-readable media further cause operations comprising:

accessing a training set of telecommunications data records comprising values indicative of network degradation; and

training the machine learning model using the training set of telecommunications data records to identify patterns in telecommunications data records corresponding to network degradation.

19 . The system of claim 18 , wherein the one or more non-transitory computer-readable media further cause operations comprising:

receiving, from a user device of the telecommunications service network, a request to identify a corrective action for mitigating network degradation experienced on the user device;

in response to the request, identifying, by the system, one or more changes that correlate with network degradation experienced on the user device; and

transmitting, by the system, instructions to the user device for mitigating the issue.

20 . The system of claim 19 , wherein the instructions comprise instructions to: (a) perform a software or hardware update, (b) reconfigure data files on the user device, or (c) manually or physically change the device.

Assignments (1)

Continuity (1)

References Cited (58)