# US 2021/0342375 A1 · System and method for automatic suggestion and or correcting of search keywords

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

- Number: US 2021/0342375 A1
- Document ID: US20210342375A1
- Status: Published application
- Kind: A1
- Application number: 16/862,640
- Filed: 2020-04-30
- Published: 2021-11-04
- Earliest claimed: 2020-04-30
- Assignees: Intuit Inc.
- Applicants: Intuit Inc.
- Inventors: Lei PEI
- Web page: https://patentgrep.com/patent/US20210342375A1

## CPC

- G06F 16/3322: {using system suggestions (G06F16/3325 takes precedence)}
- G06F 16/3329: Natural language query formulation
- G06F 16/3347: {using vector based model}
- G06F 16/353: into predefined classes
- G06F 16/9535: Search customisation based on user profiles and personalisation
- G06N 20/00: Machine learning
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

Systems and methods that may be used to automatically correct, complete and or suggest words or terms to the user of an electronic service (e.g., accounting service) while the user is entering a search keyword and or filling out a form field. The automatic correction, completion and or suggestion of words or terms are based natural language processing of historical data from a plurality of users of the electronic service.

## Claims

### Claim 1

A computer implemented method of providing automatic correction, completion or suggestions of words entered into a user interface for interacting with an electronic information system, said method being performed on a computing device, said method comprising:
inputting, from a database, collection of accounts data associated with a plurality of system users;
training a natural language processing model with the input collection of accounts data;
clustering word embeddings of user account names within the input collection of accounts data into groups using the trained natural language processing model;
inputting the clustered word embeddings into a search engine; and
using the search engine, as a user is in a process of entering a first word into a first field of the user interface, to provide a second word to automatically populate the first field.

### Claim 2 (depends on 1)

The method of claim 1, wherein training the natural language processing model with the input collection of accounts data comprises inputting the collection of accounts data into a Word2vec model to create a vector space comprising words within the input collection of accounts data, each word comprising a vector within the vector space.

### Claim 3 (depends on 2)

The method of claim 2, wherein clustering word embeddings comprises:
inputting the word embeddings;
determining a distance between word embeddings based on each word embeddings' vector; and
clustering word embeddings having a determined distance within a predetermined distance threshold.

### Claim 4 (depends on 3)

The method of claim 3, wherein the vector space comprises misspelled words and the clustered word embeddings include word embeddings of misspelled words.

### Claim 5 (depends on 1)

The method of claim 1, wherein inputting the clustered word embeddings into a search engine comprises storing the clustered word embeddings in a file accessible by the search engine.

### Claim 6 (depends on 1)

The method of claim 1, wherein the second word comprises a correct spelling of the first word and replaces the first word as the user enters the first word into the first field.

### Claim 7 (depends on 1)

The method of claim 1, wherein the second word comprises a complete spelling of the first word and completes the first word as the user enters the first word into the first field.

### Claim 8 (depends on 1)

The method of claim 1, wherein the second word comprises a synonym of the first word and replaces the first word as the user enters the first word into the first field.

### Claim 9

A system for providing automatic correction, completion or suggestions of words entered into a user interface for interacting with the system, said system providing a first service and comprising:
a first computing device connected to a database comprising collection of accounts data for a plurality of users of the first service, the first computing device configured to:
input, from the database, collection of accounts data associated with the plurality of users;
train a natural language processing model with the input collection of accounts data;
cluster word embeddings of user account names within the input collection of accounts data into groups using the trained natural language processing model;
input the clustered word embeddings into a search engine; and
use the search engine, as a user is in a process of entering a first word into a first field of the user interface, to provide a second word to automatically populate the first field.

### Claim 10 (depends on 9)

The system of claim 9, wherein the first computing device is configured to train the natural language processing model with the input collection of accounts data by inputting the collection of accounts data into a Word2vec model to create a vector space comprising words within the input collection of accounts data, each word comprising a vector within the vector space.

### Claim 11 (depends on 10)

The system of claim 10, wherein the first computing device is configured to cluster word embeddings by:
inputting the word embeddings;
determining a distance between word embeddings based on each word embeddings' vector; and
clustering word embeddings having a determined distance within a predetermined distance threshold.

### Claim 12 (depends on 11)

The system of claim 11, wherein the vector space comprises misspelled words and the clustered word embeddings include word embeddings of misspelled words.

### Claim 13 (depends on 9)

The system of claim 9, wherein the first computing device is configured to input the clustered word embeddings into a search engine by storing the clustered word embeddings in a file accessible by the search engine.

### Claim 14 (depends on 9)

The system of claim 9, wherein the second word comprises a correct spelling of the first word and wherein the first computing device is configured to replace the first word as the user enters the first word into the first field.

### Claim 15 (depends on 9)

The system of claim 9, wherein the second word comprises a complete spelling of the first word and wherein the first computing device is configured to complete the first word as the user enters the first word into the first field.

### Claim 16 (depends on 9)

The system of claim 9, wherein the second word comprises a synonym of the first word and wherein the first computing device is configured to replace the first word as the user enters the first word into the first field.
