Systems and methods for correcting incorrect product information in an electronic data catalog

Patent metadata

Field	Value
Publication number	US-11062365-B2
Application number	US-201715467854-A
Country	US
Kind code	B2
Filing date	Mar 23, 2017
Priority date	Mar 23, 2017
Publication date	Jul 13, 2021
Grant date	Jul 13, 2021

Abstract

Systems and methods including one or more processing modules and one or more non-transitory storage modules storing computing instructions configured to run on the one or more processing modules and perform acts of determining an accuracy score for existing product information using a first set of rules that compares the existing product information for the product with product information of other internal or external sources, determining if the accuracy score exceeds a predetermined accuracy threshold, automatically replacing incorrect product information in the existing product information with correct product information from the other sources if the accuracy score does not exceed the predetermined accuracy threshold, coordinating displaying of the existing product information with the correct product information replacing the incorrect product information on a webpage for the product on a website of the online retailer.

First claim

What is claimed is:

1. A system comprising:
one or more processors; and
one or more non-transitory storage devices storing computing instructions configured to run on the one or more processors and perform:
accessing existing product information for a product in a data catalog of an online retailer;
determining a respective confidence score for each respective source of one or more internal sources internal to the online retailer and one or more external sources external to the online retailer;
determining an accuracy score for the existing product information using a first set of rules that:
comprises:
a binary function that outputs a value based on a variance of an attribute of the product in the existing product information and the one or more internal sources or the one or more external sources;
a weight given to either the one or more internal sources or the one or more external sources; and
a number of the one or more internal sources or the one or more external sources; and

compares the existing product information for the product with (1) internal product information of the one or more internal sources internal to the online retailer when the respective confidence score for a respective one of the one or more internal sources is above a predetermined threshold or (2) external product information of the one or more external sources external to the online retailer when the respective confidence score for a respective one of the one or more external sources is above the predetermined threshold;

when the accuracy score exceeds a predetermined accuracy threshold, determining that product information in the existing product information is correct product information;
when the accuracy score does not exceed the predetermined accuracy threshold, automatically replacing incorrect product information in the existing product information with the correct product information from the external product information; and
coordinating displaying the existing product information with the correct product information replacing the incorrect product information on a webpage for the product on a website of the online retailer.

2. The system of claim 1 , wherein determining the accuracy score for the existing product information using the first set of rules comprises:
determining when the existing product information matches the internal product information in a database of the online retailer using machine-learned attribute extraction algorithms; and
automatically marking the existing product information when the existing product information does not match the internal product information.

3. The system of claim 2 , wherein the internal product information in the database of the online retailer comprises the internal product information (1) stored in the database of the online retailer and (2) collected by the online retailer from one or more sellers of the product or suppliers of the product.

4. The system of claim 1 , wherein determining the accuracy score for the existing product information using the first set of rules comprises:
crawling a website of one or more first external sources of the one or more external sources to collect first external product information of the external product information; and
requesting second external product information of the external product information from one or more second external sources of the one or more external sources.

5. The system of claim 1 , wherein the one or more non-transitory storage devices storing the computing instructions are further configured to run on the one or more processors and perform:
matching product identifiers of the product in:
the existing product information;
the internal product information; and
the external product information.

6. The system of claim 1 , wherein the first set of rules comprises:

Accuracy
⁢

⁢
Score
⁢

⁢

(
X
)

=

∑
i
n

⁢

wi
*
σ
⁢

⁢
x

∑
i
n

⁢
wi

,

where:
x is the attribute of the product in the existing product information;
wi is the weight given to either the one or more internal sources or the one or more external sources;
σx is the binary function that outputs the value based on the variance of the attribute in the existing product information and the one or more internal sources or the one or more external sources; and
n is the number of the one or more internal sources or the one or more external sources.

7. The system of claim 6 , wherein:
σx is also equal to a length of a longest matching substring of the one or more internal sources or the one or more external sources divided by a length of a substring in the existing product information.

8. The system of claim 1 , wherein:
determining the accuracy score for the existing product information using the first set of rules comprises:
determining when the existing product information matches the internal product information in a database of the online retailer using machine-learned attribute extraction algorithms;
automatically marking the existing product information when the existing product information does not match the internal product information;
crawling a website of one or more first external sources of the one or more external sources to collect first external product information of the external product information; and
requesting second external product information of the external product information from one or more second external sources of the one or more external sources;

the internal product information in the database of the online retailer comprises the internal product information (1) stored in the database of the online retailer and (2) collected by the online retailer from one or more sellers of the product or suppliers of the product;
the one or more non-transitory storage devices storing the computing instructions are further configured to run on the one or more processors and perform:
matching product identifiers of the product in the existing product information, the internal product information, and the external product information;

the first set of rules comprises:

Accuracy
⁢

⁢
Score
⁢

⁢

(
X
)

Assignees

• Wal Mart Stores Inc
• Walmart Apollo Llc

Inventors

• Siddiqui Omar Saeed
• Deep Rohit
• Nafar Zuzar Fakhruddin

Classifications

• G06F16/215

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

• G06Q30/0641

  utilising user interfaces specially adapted for shopping · CPC title

• G06N20/00

  Machine learning · CPC title

• G06F16/24578

  using ranking · CPC title

• G06F16/248

  Presentation of query results · CPC title