# US 2026/0065100 A1 · Algorithm selector for profiling application usage based on network signals

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0065100 A1
- Document ID: US20260065100A1
- Status: Published application
- Kind: A1
- Application number: 19/384,287
- Filed: 2025-11-10
- Published: 2026-03-05
- Earliest claimed: 2021-11-30
- Assignees: T-Mobile USA, Inc.
- Applicants: T-Mobile USA, Inc.
- Inventors: Stewart Renehan; Jeremy Angel Evan
- Web page: https://patentgrep.com/patent/US20260065100A1

## CPC

- G06N 5/04: Inference or reasoning models
- G06N 3/044: Recurrent networks, e.g. Hopfield networks
- G06N 3/045: Combinations of networks
- G06N 3/08: Learning methods
- G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
- G06N 5/022: Knowledge engineering; Knowledge acquisition
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- H04L 41/16: using machine learning or artificial intelligence
- H04L 43/026: using flow identification
- H04L 43/04: Processing captured monitoring data, e.g. for logfile generation
- H04L 67/535: {Tracking the activity of the user (network monitoring arrangements H04L43/00; recording of computer activity G06F11/34)}
- G06N 20/00: Machine learning
- G06N 20/10: using kernel methods, e.g. support vector machines [SVM]
- G06N 20/20: Ensemble learning
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

The disclosed technique incudes a method for dynamically selecting an algorithm that predicts usage of a mobile application based on network signals. In one example, an application profile includes known network signals and designates a best predictive algorithm. When network signals of user devices are subsequently captured, the best algorithms of sufficiently matching profiles are used to estimate application usage. As such, for example, the popularity of a particular application or relationships among applications can be determined for managing the network or for commercial purposes.

## Claims

### Claim 1

A computer-readable storage medium, excluding transitory signals and carrying instructions, which, when executed by at least one data processor of a system, cause the system to:
process Server Name Indication (SNI) signals of a first application using multiple different machine learning (ML) algorithms to determine respective predictions of a source of the SNI signals,
wherein the SNI signals of the first application are communicated by a first wireless user device over a wireless telecommunications network and indicates a host for the first application;
determine, from the multiple different ML algorithms, one or more preferred ML algorithms based on a confidence of the respective predictions;
determine that subsequent SNI signals and the SNI signals satisfy a similarity threshold; and
based on determining that the subsequent SNI signals and the SNI signals satisfy the similarity threshold, process the subsequent SNI signals using the one or more preferred ML algorithms.

### Claim 2 (depends on 1)

The computer-readable storage medium of claim 1, wherein to determine the one or more preferred ML algorithms comprises causing the system to:
determine a classifier algorithm configured to predict whether the first application is utilized by the first wireless user device; and
determine a regression algorithm configured to predict a duration that the first application was utilized by the first wireless user device.

### Claim 3 (depends on 1)

The computer-readable storage medium of claim 1, wherein the system is further caused to:
prior to processing the SNI signals using the multiple different ML algorithms, analyze the SNI signals to determine that the SNI signals originate from at least one application; and
in response to determining that the SNI signals originate from the at least one application, process the SNI signals using the multiple different ML algorithms.

### Claim 4 (depends on 1)

The computer-readable storage medium of claim 1, wherein the system is further caused to:
prior to processing the subsequent SNI signals using the one or more preferred ML algorithms, analyze the subsequent SNI signals to determine that the subsequent SNI signals originate from at least one application; and
in response to determining that the subsequent SNI signals originate from the at least one application, process the subsequent SNI signals using the one or more preferred ML algorithms.

### Claim 5 (depends on 1)

The computer-readable storage medium of claim 1, wherein the system is further caused to:
receive additional SNI signals of the first application;
process the additional SNI signals of the first application using one or more of the multiple different ML algorithms or additional ML algorithms to determine respective predictions of a source of the additional SNI signals;
determine that a confidence of a respective prediction of an additional ML algorithm from the one or more multiple different ML algorithms or the additional ML algorithms is higher than a confidence of the one or more preferred ML algorithms; and
process future SNI signals of the first application using the additional ML algorithm.

### Claim 6 (depends on 1)

The computer-readable storage medium of claim 1, wherein the system is further caused to:
monitor utilization of the first application using the one or more preferred ML algorithms; and
determine a popularity of the first application based on the utilization of the first application.

### Claim 7 (depends on 1)

The computer-readable storage medium of claim 1, wherein the system is further caused to:
monitor utilization of the first application by the first wireless user device using the one or more preferred ML algorithms; and
determine a persona for a user of the first wireless user device based on the utilization of the first application by the first wireless user device.

### Claim 8 (depends on 1)

The computer-readable storage medium of claim 1, wherein the system is further caused to:
monitor utilization of the first application using the one or more preferred ML algorithms; and
determine one or more second applications that have a usage pattern correlated with the utilization of the first application based on the utilization of the first application.

### Claim 9 (depends on 1)

The computer-readable storage medium of claim 1, wherein the subsequent SNI signals are communicated by a second wireless user device different from the first wireless user device.

### Claim 10

A system comprising:
at least one hardware processor; and
at least one non-transitory memory storing instructions, which, when executed by the at least one hardware processor, cause the system to:
process Server Name Indication (SNI) signals of a first application using multiple different machine learning (ML) algorithms to determine respective predictions of a source of the SNI signals,
wherein the SNI signals of the first application are communicated by a first wireless user device over a wireless telecommunications network and indicates a host for the first application;
determine, from the multiple different ML algorithms, one or more preferred ML algorithms based on a confidence of the respective predictions;
determine that subsequent SNI signals and the SNI signals satisfy a similarity threshold; and
based on determining that the subsequent SNI signals and the SNI signals satisfy the similarity threshold, process the subsequent SNI signals using the one or more preferred ML algorithms.

### Claim 11 (depends on 10)

The system of claim 10, wherein the system is further caused to:
receive additional SNI signals of the first application;
process the additional SNI signals of the first application using one or more of the multiple different ML algorithms or additional ML algorithms to determine respective predictions of a source of the additional SNI signals;
determine that a confidence of a respective prediction of an additional ML algorithm from the one or more multiple different ML algorithms or the additional ML algorithms is higher than a confidence of the one or more preferred ML algorithms; and
process future SNI signals of the first application using the additional ML algorithm.

### Claim 12 (depends on 10)

The system of claim 10, wherein the system is further caused to:
monitor utilization of the first application using the one or more preferred ML algorithms; and
determine a popularity of the first application based on the utilization of the first application.

### Claim 13 (depends on 10)

The system of claim 10, wherein the system is further caused to:
monitor utilization of the first application by the first wireless user device using the one or more preferred ML algorithms; and
determine a persona for a user of the first wireless user device based on the utilization of the first application by the first wireless user device.

### Claim 14 (depends on 10)

The system of claim 10, wherein the system is further caused to:
monitor utilization of the first application using the one or more preferred ML algorithms; and
determine one or more second applications that have a usage pattern correlated with the utilization of the first application based on the utilization of the first application.

### Claim 15 (depends on 10)

The system of claim 10, wherein the subsequent SNI signals are communicated by a second wireless user device different from the first wireless user device.

### Claim 16

A method comprising:
processing Server Name Indication (SNI) signals of a first application using multiple different machine learning (ML) algorithms to determine respective predictions of a source of the SNI signals,
wherein the SNI signals of the first application are communicated by a first wireless user device over a wireless telecommunications network and indicates a host for the first application;
determining, from the multiple different ML algorithms, one or more preferred ML algorithms based on a confidence of the respective predictions;
determining that subsequent SNI signals and the SNI signals satisfy a similarity threshold; and
based on determining that the subsequent SNI signals and the SNI signals satisfy the similarity threshold, processing the subsequent SNI signals using the one or more preferred ML algorithms.

### Claim 17 (depends on 16)

The method of claim 16, further comprising:
receiving additional SNI signals of the first application;
processing the additional SNI signals of the first application using one or more of the multiple different ML algorithms or additional ML algorithms to determine respective predictions of a source of the additional SNI signals;
determining that a confidence of a respective prediction of an additional ML algorithm from the one or more multiple different ML algorithms or the additional ML algorithms is higher than a confidence of the one or more preferred ML algorithms; and
processing future SNI signals of the first application using the additional ML algorithm.

### Claim 18 (depends on 16)

The method of claim 16, further comprising:
monitoring utilization of the first application using the one or more preferred ML algorithms; and
determining a popularity of the first application based on the utilization of the first application.

### Claim 19 (depends on 16)

The method of claim 16, further comprising:
monitoring utilization of the first application using the one or more preferred ML algorithms; and
determining one or more second applications that have a usage pattern correlated with the utilization of the first application based on the utilization of the first application.

### Claim 20 (depends on 16)

The method of claim 16, wherein the subsequent SNI signals are communicated by a second wireless user device different from the first wireless user device.
