# US 12,118,442 B2 · Systems and methods for automatically determining a preferred communication channel for messaging

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

- Number: US 12,118,442 B2
- Document ID: US12118442B2
- Status: Granted
- Kind: B2
- Application number: 18/103,398
- Filed: 2023-01-30
- Published: 2023-08-10
- Granted: 2024-10-15
- Earliest claimed: 2019-02-27
- Assignees: Walmart Apollo, LLC
- Applicants: Walmart Apollo, LLC
- Inventors: Sayan Maity; Wei Shen
- Web page: https://patentgrep.com/patent/US12118442B2

## CPC

- G06N 20/00: Machine learning
- H04L 51/18: Commands or executable codes
- G06F 17/18: for evaluating statistical data {, e.g. average values, frequency distributions, probability functions, regression analysis (forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04)}
- G06N 7/00: Computing arrangements based on specific mathematical models
- G06N 20/20: Ensemble learning
- G06Q 30/0203: Market surveys; Market polls
- H04L 51/216: Handling conversation history, e.g. grouping of messages in sessions or threads
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- 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)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR

## Abstract

A method including determining a preferred communication channel for a user by: (a) determining a respective normalized propensity score for each subscribed communication channel to which the user subscribes, based at least in part on a respective propensity score for the each subscribed communication channel; and (b) determining the preferred communication channel from among subscribed communication channels based on the respective normalized propensity score. The method further can include transmitting, via the preferred communication channel, a message to a user device of the user. Other embodiments are disclosed.

## Claims

### Claim 1

A system comprising:
one or more processors; and
one or more non-transitory computer-readable media storing computing instructions configured to, when run on the one or more processors, cause the one or more processors to perform:
determining a preferred communication channel for a user by:
determining a respective normalized propensity score for each subscribed communication channel to which the user subscribes, based at least in part on a respective propensity score for each subscribed communication channel; and
determining the preferred communication channel from among subscribed communication channels based on the respective normalized propensity score; and
transmitting, via the preferred communication channel, a message to a user device of the user.

### Claim 2 (depends on 1)

The system in claim 1, wherein:
determining the respective normalized propensity score for each subscribed communication channel further comprises:
validating a normality of the respective propensity score for each subscribed communication channel;
when the normality of the respective propensity score is validated, setting the respective normalized propensity score for each subscribed communication channel as equal to the respective propensity score; and
when the normality of the respective propensity score is not validated, determining the respective normalized propensity score by normalizing the respective propensity score for each subscribed communication channel.

### Claim 3 (depends on 2)

The system in claim 2, wherein:
normalizing the respective propensity score for each subscribed communication channel further comprises using one of:
a log transformation technique;
a Platt scaling technique;
a min-max normalization technique;
a Z-score normalization technique; or
an isotonic regression technique.

### Claim 4 (depends on 2)

The system in claim 2, wherein:
determining the respective normalized propensity score for each subscribed communication channel further comprises:
after normalizing the respective propensity score, re-validating the normality of the respective propensity score; and
after re-validating the normality of the respective propensity score, when the normality of the respective propensity score is not validated, repeating, up to a predetermined count unless the normality of the respective propensity score is validated:
re-normalizing the respective propensity score; and
re-validating the normality of the respective propensity score.

### Claim 5 (depends on 4)

The system in claim 4, wherein:
a normalization technique for re-normalizing the respective propensity score for each subscribed communication channel is different from a normalization technique for normalizing the respective propensity score.

### Claim 6 (depends on 1)

The system in claim 1, wherein:
determining the respective normalized propensity score for each subscribed communication channel further comprises:
extracting at least one feature vector from subscription data for the user;
calculating the respective propensity score for each subscribed communication channel using the at least one feature vector; and
determining the respective normalized propensity score by normalizing the respective propensity score.

### Claim 7 (depends on 6)

The system in claim 6, wherein:
the computing instructions are further configured to, when run on the one or more processors, cause the one or more processors to perform:
after calculating the respective propensity score for each subscribed communication channel, calibrating the respective normalized propensity score using a prior correction technique.

### Claim 8 (depends on 7)

The system in claim 7, wherein:
the prior correction technique comprises an equation of:
=
ln
[
(
1
-
τ
τ
)
⁢
(
y
¯
1
-
y
¯
)
]
,
where comprises an intercept of a logistic regression model;
τ comprises a fraction of ones in a population; and
y comprises a fraction of ones in a sample.

### Claim 9 (depends on 1)

The system in claim 1, wherein:
determining the preferred communication channel further comprises:
defining score-based categories for the subscribed communication channels, each of the score-based categories being empty initially;
assigning the respective normalized propensity score to a respective score category of the score-based categories;
when a single channel of the subscribed communication channels exists in a highest score category of one or more non-empty categories of the score-based categories, setting the preferred communication channel to the single channel; and
when the single channel in the highest score category does not exist:
repeating, until the single channel is found:
re-defining the score-based categories into finer granularity; and
re-assigning the respective normalized propensity score into the respective score category of the score-based categories; and
setting the preferred communication channel to the single channel.

### Claim 10 (depends on 1)

The system in claim 1, wherein:
the computing instructions are further configured to, when run on the one or more processors, cause the one or more processors to perform:
extracting a subscription history and an engagement history from subscription data for the user; and
transmitting, via the preferred communication channel, the message to the user device of the user comprises, when (a) (i) the subscription history is short, and (ii) no message open record exists in the engagement history; or (b) (i) the subscription history is not short, and (ii) no message open record exists in a prior predetermined period of the engagement history, transmitting, via each subscribed communication channel, the message to the user device of the user.

### Claim 11

A method being implemented via execution of computing instructions configured to run at one or more processors and configured to be stored at non-transitory computer-readable media, the method comprising:
determining a preferred communication channel for a user by:
determining a respective normalized propensity score for each subscribed communication channel to which the user subscribes, based at least in part on a respective propensity score for each subscribed communication channel; and
determining the preferred communication channel from among subscribed communication channels based on the respective normalized propensity score; and
transmitting, via the preferred communication channel, a message to a user device of the user.

### Claim 12 (depends on 11)

The method in claim 11, wherein:
determining the respective normalized propensity score for each subscribed communication channel further comprises:
validating a normality of the respective propensity score for each subscribed communication channel;
when the normality of the respective propensity score is validated, setting the respective normalized propensity score for each subscribed communication channel as equal to the respective propensity score; and
when the normality of the respective propensity score is not validated, determining the respective normalized propensity score by normalizing the respective propensity score for each subscribed communication channel.

### Claim 13 (depends on 12)

The method in claim 12, wherein:
normalizing the respective propensity score for each subscribed communication channel further comprises using one of:
a log transformation technique;
a Platt scaling technique;
a min-max normalization technique;
a Z-score normalization technique; or
an isotonic regression technique.

### Claim 14 (depends on 12)

The method in claim 12, wherein:
determining the respective normalized propensity score for each subscribed communication channel further comprises:
after normalizing the respective propensity score, re-validating the normality of the respective propensity score; and
after re-validating the normality of the respective propensity score, when the normality of the respective propensity score is not validated, repeating, up to a predetermined count unless the normality of the respective propensity score is validated:
re-normalizing the respective propensity score; and
re-validating the normality of the respective propensity score.

### Claim 15 (depends on 14)

The method in claim 14, wherein:
a normalization technique for re-normalizing the respective propensity score for each subscribed communication channel is different from a normalization technique for normalizing the respective propensity score.

### Claim 16 (depends on 11)

The method in claim 11, wherein:
determining the respective normalized propensity score for each subscribed communication channel further comprises:
extracting at least one feature vector from subscription data for the user;
calculating the respective propensity score for each subscribed communication channel using the at least one feature vector; and
determining the respective normalized propensity score by normalizing the respective propensity score.

### Claim 17 (depends on 16)

The method in claim 16 further comprising:
after calculating the respective propensity score for each subscribed communication channel, calibrating the respective normalized propensity score using a prior correction technique.

### Claim 18 (depends on 17)

The method in claim 17, wherein:
the prior correction technique comprises an equation of:
=
ln
[
(
1
-
τ
τ
)
⁢
(
y
¯
1
-
y
¯
)
]
,
where comprises an intercept of a logistic regression model;
τ comprises a fraction of ones in a population; and
y comprises a fraction of ones in a sample.

### Claim 19 (depends on 11)

The method in claim 11, wherein:
determining the preferred communication channel further comprises:
defining score-based categories for the subscribed communication channels, each of the score-based categories being empty initially;
assigning the respective normalized propensity score to a respective score category of the score-based categories;
when a single channel of the subscribed communication channels exists in a highest score category of one or more non-empty categories of the score-based categories, setting the preferred communication channel to the single channel; and
when the single channel in the highest score category does not exist:
repeating, until the single channel is found:
re-defining the score-based categories into finer granularity; and
re-assigning the respective normalized propensity score into the respective score category of the score-based categories; and
setting the preferred communication channel to the single channel.

### Claim 20 (depends on 11)

The method in claim 11, wherein:
the method further comprises:
extracting a subscription history and an engagement history from subscription data for the user; and
transmitting, via the preferred communication channel, the message to the user device of the user comprises, when (a) (i) the subscription history is short, and (ii) no message open record exists in the engagement history; or (b) (i) the subscription history is not short, and (ii) no message open record exists in a prior predetermined period of the engagement history, transmitting, via each subscribed communication channel, the message to the user device of the user.