# US 12,111,849 B2 · Managing data processing efficiency, and applications thereof

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

- Number: US 12,111,849 B2
- Document ID: US12111849B2
- Status: Granted
- Kind: B2
- Application number: 17/699,358
- Filed: 2022-03-21
- Published: 2022-10-06
- Granted: 2024-10-08
- Earliest claimed: 2016-03-16
- Assignees: VEDA Data Solutions, Inc.
- Applicants: VEDA Data Solutions, Inc.
- Inventors: Robert Raymond Lindner
- Web page: https://patentgrep.com/patent/US12111849B2

## CPC

- G06F 16/285: {Clustering or classification}
- G06F 16/215: Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
- G06F 16/2465: {Query processing support for facilitating data mining operations in structured databases}
- G06F 40/205: Parsing
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

Disclosed herein are system, method, and computer program product embodiments for linking data records in memory. The system, method, and computer program product includes accessing a first record stored in memory, the first record holding information describing a first person and accessing at least one additional record stored in memory, the additional records holding information describing additional persons. The method continues by parsing the information of the first record and additional record and assigning the parsed information to predefined categories within the respective records. After assigning the information into categories, a similarity score between categorical information in the first record and categorical information of additional records is determined. A category of an additional record is then modified based on the similarity score, so the additional record is associated with the first person.

## Claims

### Claim 1

A computer-implemented method, comprising:
accessing, by one or more processors, a data record describing an individual;
assigning, by the one or more processors, the data record to a predetermined category;
comparing, by the one or more processors, the categorized data record against other data records stored in a database using a pair-wise function in order to determine whether the categorized data record should be linked, grouped, or modified to mirror an identity described by a separate data record of the other data records;
entering, by the one or more processors and into a training system, the categorized data record;
training, by the one or more processors, the training system using the categorized data record to find further individuals possessing similar demographic data as the individual to which the data record belongs;
finding, by the one or more processors, the further individuals by accessing further data records of individuals and processing the further data records using the trained training system; and
predicting, by the one or more processors, future behaviors of the further individuals based on an order of similarity between data records of the further individuals and the individual.

### Claim 2 (depends on 1)

The method of claim 1, wherein the data record comprises: income data, consumer data, web-browsing data, or an individual's mortgage history.

### Claim 3 (depends on 2)

The method of claim 2, further comprising linking, grouping, or modifying, by the one or more processors, the categorized data record to the separate data record when the pair-wise function results in a similarity score exceeding a predetermined threshold.

### Claim 4

A non-transitory computer readable medium storing instructions, that when executed by one or more processors, cause the one or more processors to perform operations comprising:
accessing a data record describing an individual;
assigning the data record to a predetermined category;
comparing the categorized data record against other data records stored in a database using a pair-wise function in order to determine whether the categorized data record should be linked, grouped, or modified to mirror the identity described by a separate data record of the other data records;
entering, into a training system, the categorized data record;
training the training system using the categorized data record to find further individuals possessing similar demographic data as an individual to which the data record belongs;
finding the further individuals by accessing further data records of individuals and processing the further data records using the trained training system; and
predicting future behaviors of the further individuals based on an order of similarity between data records of the further individuals and the individual.

### Claim 5 (depends on 4)

The non-transitory computer readable medium of claim 4, wherein the data record comprises: income data, consumer data, web-browsing data, or an individual's mortgage history.

### Claim 6 (depends on 4)

The non-transitory computer readable medium of claim 4, wherein the operations further comprise linking, grouping, or modifying the categorized data record to the separate data record when the pair-wise function results in a similarity score exceeding a predetermined threshold.

### Claim 7

A computing system, comprising:
a memory storing instructions;
one or more processors, coupled to the memory, configured to process the stored instructions to:
access a data record describing an individual;
assign the data record to a predetermined category;
compare the categorized data record against other data records stored in a database using a pair-wise function in order to determine whether the categorized data record should be linked, grouped, or modified to mirror an identity described by a separate data record of the other data records;
enter, into a training system, the categorized data record; and
train the training system using the categorized data record to find further individuals possessing similar demographic data as the individual to which the data record belongs;
find the further individuals by accessing further data records of individuals and processing the further data records using the trained training system;
predict future behaviors of the further individuals based on an order of similarity between data records of the further individuals and the individual.

### Claim 8 (depends on 7)

The computing system of claim 7, wherein the data record comprises: income data, consumer data, web-browsing data, or an individual's mortgage history.

### Claim 9 (depends on 7)

The computing system of claim 7, wherein the one or more processors are further configured to:
link, group, or modify the categorized data record to the separate data record when the pair-wise function results in a similarity score exceeding a predetermined threshold.