# US 2021/0004422 A1 · User electronic message system

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

- Number: US 2021/0004422 A1
- Document ID: US20210004422A1
- Status: Published application
- Kind: A1
- Application number: 16/503,163
- Filed: 2019-07-03
- Published: 2021-01-07
- Earliest claimed: 2019-07-03
- Assignees: Airbnb, Inc.
- Applicants: Airbnb, Inc.
- Inventors: Zhentao Sun; Dawei Wang; Sean Odonnell; Noah Martin; Andriy Sergiyenko; Greg Thompson; Shike Mei
- Web page: https://patentgrep.com/patent/US20210004422A1

## CPC

- G06F 16/9535: Search customisation based on user profiles and personalisation
- G06F 16/9538: Presentation of query results
- G06N 20/00: Machine learning
- H04L 51/212: using filtering or selective blocking
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- 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

Network site users can be selected to receive a communication based on a network site event, such as incomplete registration. A hybrid user interaction machine learning scheme can select a portion of the selected users based on user interaction estimates and network sampling data. The electronic document sent to the users can have portions that undergo two-pass ranking for ordering of content items to be included in the electronic document, such as an email.

## Claims

### Claim 1

A method comprising:
identifying, using one or more processors of a machine, a plurality of users of a network site;
generating, using a machine learning scheme of the network site, values of a first user interaction type for the plurality of users, each of the values of the first interaction type indicating a likelihood that a given user of the plurality of users interacts with an electronic document of the network site;
partitioning the values of the first interaction type into ranges;
live network testing, on the network site, a sample of users from each of the ranges to identify one or more top ranges that comprise users that perform an elevated amount of a second interaction type with the network site, the second user interaction type being a different user interaction with the network site than the first user interaction type;
storing one or more users that are in the one or more top ranges based on the one or more users having first interaction type values, generated by the machine learning scheme, that are in the one or more top ranges; and
transmitting, by the network site, an electronic document to the one or more users.

### Claim 2 (depends on 1)

The method of claim 1, wherein identifying the plurality of users comprises:
determining that the plurality of users correspond to a preconfigured network site event.

### Claim 3 (depends on 2)

The method of claim 2, wherein the preconfigured network site event is abandonment of a network site navigation path.

### Claim 4 (depends on 3)

The method of claim 3, wherein the network site navigation path is a network site transaction.

### Claim 5 (depends on 1)

The method of claim 1, wherein the first user interaction type is an open electronic document interaction.

### Claim 6 (depends on 1)

The method of claim 1, further comprising:
training the machine learning scheme using training data comprising historical network site document data.

### Claim 7 (depends on 6)

The method of claim 6, wherein the historical network site data comprises attribute values describing attributes of previous electronic documents of the network site.

### Claim 8 (depends on 1)

The method of claim 1, wherein the machine learning scheme is a decision tree classifier machine learning scheme.

### Claim 9 (depends on 1)

The method of claim 1, wherein the values of the first interaction type are partitioned into ranges using a grouping machine learning scheme.

### Claim 10 (depends on 1)

The method of claim 1, wherein the values of the first interaction type are partitioned into ranges using a set of pre-configured ranges.

### Claim 11 (depends on 1)

The method, of claim 1, further comprising:
generating, using a first ranking machine learning scheme, an unpopulated content scheme ranking data set that ranks a plurality of unpopulated content schemes;
generating, for each of the plurality of users, a plurality of populated content schemes by populating the plurality of unpopulated content schemes with network site content items that match the user profile data of each user;
generating, using a second ranking machine learning scheme, a populated content scheme ranking data set that ranks the plurality of populated content schemes, the second ranking machine learning scheme receiving input data that includes output of the first ranking machine learning scheme that is trained to rank unpopulated content schemes;
generating, for each of the plurality of users, the electronic document using one or more top ranked populated content schemes.

### Claim 12 (depends on 11)

The method of claim 11, wherein the unpopulated content schemes are network item retrieval schemes that order network site items in a user interface window.

### Claim 13 (depends on 11)

The method of claim 11, wherein the unpopulated content schemes are populated by selecting network items according to each content scheme.

### Claim 14

A system comprising:
one or more processors of a machine; and
a memory storing instructions that, when executed by the one or more processors, cause the machine to perform operations comprising:
identifying, using one or more processors of a machine, a plurality of users of a network site;
generating, using a machine learning scheme of the network site, values of a first user interaction type for the plurality of users, each of the values of the first interaction type indicating a likelihood that a given user of the plurality of users interacts with an electronic document of the network site;
partitioning the values of the first interaction type into ranges;
live network testing, on the network site, a sample of users from each of the ranges to identify one or more top ranges that comprise users that perform an elevated amount of a second interaction type with the network site, the second user interaction type being a different user interaction with the network site than the first user interaction type;
storing one or more users that are in the one or more top ranges based on the one or more users having first interaction type values, generated by the machine learning scheme, that are in the one or more top ranges; and
transmitting, by the network site, an electronic document to the one or more users.

### Claim 15 (depends on 14)

The system of claim 14, wherein identifying the plurality of users comprises: determining that the plurality of users correspond to a preconfigured network site event.

### Claim 16 (depends on 15)

The system of claim 15, wherein the preconfigured network site event is abandonment of a network site navigation path.

### Claim 17 (depends on 16)

The system of claim 16, wherein the network site navigation path is a network site transaction.

### Claim 18 (depends on 14)

The system of claim 14, wherein the first user interaction type is an open electronic document interaction.

### Claim 19

A machine-readable storage device embodying instructions that, when executed by a machine, cause the machine to perform operations comprising:
identifying, using one or more processors of a machine, a plurality of users of a network site;
generating, using a machine learning scheme of the network site, values of a first user interaction type for the plurality of users, each of the values of the first interaction type indicating a likelihood that a given user of the plurality of users interacts with an electronic document of the network site;
partitioning the values of the first interaction type into ranges;
live network testing, on the network site, a sample of users from each of the ranges to identify one or more top ranges that comprise users that perform an elevated amount of a second interaction type with the network site, the second user interaction type being a different user interaction with the network site than the first user interaction type;
storing one or more users that are in the one or more top ranges based on the one or more users having first interaction type values, generated by the machine learning scheme, that are in the one or more top ranges; and
transmitting, by the network site, an electronic document to the one or more users.

### Claim 20 (depends on 14)

The machine-readable storage device of claim 14, wherein identifying the plurality of users comprises: determining that the plurality of users correspond to a preconfigured network site event.
