Systems and methods for asset identification string error detection and correction

Patent metadata

Field	Value
Publication number	US-11995053-B2
Application number	US-202217854771-A
Country	US
Kind code	B2
Filing date	Jun 30, 2022
Priority date	Jun 30, 2022
Publication date	May 28, 2024
Grant date	May 28, 2024

Abstract

A method for asset identification string error detection can include receiving an asset identification string corresponding to an asset and aggregating data associated with the asset from multiple sources. The method can include creating one or more probabilistic tables linking the aggregated data to the asset and determining a likelihood that at least a portion of the asset identification string is correct based on the probabilistic tables. The method can indicate that the asset identification string includes an error when the likelihood is less than a selected threshold and otherwise indicate that the asset identification string is correct when the likelihood is greater than or equal to the selected threshold.

First claim

What is claimed is:

1. One or more memory devices storing instructions thereon for asset identification string error detection, the instructions, when executing by one or more processors, configured to cause the one or more processors to perform operations comprising:
receiving an asset identification string associated with an asset;
aggregating data associated with the asset from multiple source computing systems, the multiple source computing systems comprising two or more of asset owner data, asset dealer data, asset manufacturing data, asset sales data, and asset use data;
creating one or more probabilistic tables linking the aggregated data to the asset;
using a computer-based model,
using the one or more probabilistic tables, determining a likelihood that at least a portion of the asset identification string indicates that an association between the asset identification string and the asset is correct;
indicating that the association between the asset identification string and the asset is incorrect when the likelihood is less than a tunable threshold;
indicating that the association between the asset identification string and the asset is correct when the likelihood is greater than or equal to the tunable threshold;
upon detecting that the association between the asset identification string and the asset is incorrect, replacing the portion with a replacement string when the replacement string has a probability based on the tables greater than the tunable threshold or otherwise generating at least one replacement string recommendation; and
updating the tunable threshold of the computer-based model to integrate a verification result for the at least one replacement string recommendation.

2. The one or more memory devices of claim 1 , wherein the asset identification string includes at least a prefix substring and a root substring.

3. The one or more memory devices of claim 2 , wherein the portion of the asset identification string is the prefix substring.

4. The one or more memory devices of claim 2 , wherein the portion of the asset identification string is the root substring.

5. The one or more memory devices of claim 1 , the operations further comprising presenting the replacement string or replacement string recommendations to a user for verification.

6. A method for asset identification string error detection and correction, the method comprising:
receiving an asset identification string associated with an asset;
verifying that the asset identification string is valid and correctly formatted;
aggregating data associated with the asset from multiple source computing systems, the multiple source computing systems comprising two or more of asset owner data, asset dealer data, asset manufacturing data, asset sales data, and asset use data;
creating one or more probabilistic tables linking the aggregated data to the asset;
using a computer-based model,
using the one or more probabilistic tables, determining a likelihood that at least a portion of the asset identification string indicates that an association between the asset identification string and the asset is correct;
indicating that the association between the asset identification string and the asset is incorrect when the likelihood is less than a tunable threshold;
indicating that the association between the asset identification string and the asset is correct when the likelihood is greater than or equal to the tunable threshold;
upon detecting that the association between the asset identification string and the asset is incorrect, replacing the portion with a replacement string when the replacement string has a probability based on the tables greater than the tunable threshold or otherwise generating at least one replacement string recommendation; and
updating the tunable threshold of the computer-based model to integrate a verification result for the at least one replacement string recommendation.

7. The method of claim 6 , wherein the asset identification string includes at least a prefix substring and a root substring.

8. The method of claim 7 , wherein the portion of the asset identification string is the prefix substring.

9. The method of claim 7 , wherein the portion of the asset identification string is the root substring.

10. The method of claim 6 , further comprising presenting the replacement string or replacement string recommendations to a user for verification.

11. A system for asset identification string error detection, comprising:
one or more processors; and
one or more memory devices having stored thereon instructions that when executed by the one or more processors cause the one or more processors to:
receive an asset identification string associated with an asset;
aggregate data associated with the asset from multiple source computing systems, the multiple source computing systems comprising two or more of asset owner data, asset dealer data, asset manufacturing data, asset sales data, and asset use data;
create one or more probabilistic tables linking the aggregated data to the asset;
using a computer-based model,
using the one or more probabilistic tables, determine a likelihood that at least a portion of the asset identification string indicates that an association between the asset identification string and the asset is correct;
indicate that the association between the asset identification string and the asset is incorrect includes an error when the likelihood is less than a tunable threshold;
indicate that the association between the asset identification string and the asset is correct when the likelihood is greater than or equal to the tunable threshold;
upon detecting that the association between the asset identification string and the asset is incorrect, replace the portion with a replacement string when the replacement string has a probability based on the tables greater than the tunable threshold or otherwise generating at least one replacement string recommendation; and
update the tunable threshold of the computer-based model to integrate a verification result for the at least one replacement string recommendation.

12. The system of claim 11 , wherein the asset identification string includes at least a prefix substring and a root substring.

13. The system of claim 12 , wherein the portion of the asset identification string is the prefix substring.

14. The system of claim 12 , wherein the portion of the asset identification string is the root substring.

15. The system of claim 11 , further comprising presenting the replacement string or replacement string recommendations to a user for verification.

Assignees

• Caterpillar Inc

Inventors

• Tafaghodi Khajavi Navid
• Mason James
• Radakovic Daniela

Classifications

• G06F16/215 Primary

  Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors · CPC title

• G06Q10/10 Primary

  Office automation; Time management · CPC title

• G06Q10/087

  Inventory or stock management, e.g. order filling, procurement or balancing against orders · CPC title