# US 8,280,894 B2 · Method and system for maintaining item authority

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

- Number: US 8,280,894 B2
- Document ID: US8280894B2
- Status: Granted
- Kind: B2
- Application number: 10/350,144
- Filed: 2003-01-22
- Published: 2004-07-22
- Granted: 2012-10-02
- Earliest claimed: 2003-01-22
- Assignees: Amazon Technologies, Inc.
- Inventors: Shawn Bohn; Anmol Paralkar; Anuvrata Arora; Nicholas Bicknell
- Web page: https://patentgrep.com/patent/US8280894B2

## CPC

- G06Q 30/02: Marketing; Price estimation or determination; Fundraising
- G06Q 30/0601: Electronic shopping [e-shopping]
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR

## Abstract

An item authority system is provided. The item authority system uses rules to identify item definitions that match or potentially match an item description. When a unique match is found, then the item authority system may indicate that the item description describes the same item as the item definition. If multiple matches or only potential matches are identified, then the item authority system may allow a user to manually indicate which item definition matches.

## Claims

### Claim 1

A method in a computer system for identifying an item definition that matches an item description, the item definition and item description having attributes with values, the method comprising:
providing one or more rules that specify how to generate a similarity score based on similarity between the values of the attributes of an item definition and an item description;
for each of a plurality of item definitions, generating a similarity score for the item definition and the item description in accordance with the rules, wherein a score is assigned to the attributes of the item definition and the item description, and scores of the attributes are aggregated to derive the similarity score; and
selecting the item definition whose generated similarity score indicates it is most similar to the item description as the matching item definition.

### Claim 2 (depends on 1)

The method of claim 1 wherein a rule indicates how to generate a similarity score based on similarity of the values of multiple attributes.

### Claim 3 (depends on 1)

The method of claim 1 wherein a rule includes one or more filters, each filter specifying how to generate a filter similarity score, wherein the similarity score for an item definition is the filter similarity score that indicates the item definition is most similar to the item description.

### Claim 4 (depends on 3)

The method of claim 3 wherein a filter specifies an attribute scoring technique for one or more attributes that generates an attribute similarity score for that attribute and specifies how to combine the generated attribute similarity scores to generate a filter similarity score.

### Claim 5 (depends on 4)

The method of claim 4 wherein an attribute scoring technique specifies a minimum score for the attribute, wherein, when the minimum score is not met, the filter similarity score for the filter is set to indicate no similarity.

### Claim 6 (depends on 4)

The method of claim 4 wherein an attribute scoring technique specifies that, when the attribute is not present in either the item definition or the item description, the filter similarity score is set to indicate no similarity.

### Claim 7 (depends on 4)

The method of claim 4 wherein an attribute scoring technique specifies that, when a value for the attribute is not valid in either the item definition or the item description, the filter similarity score is set to indicate no similarity.

### Claim 8 (depends on 4)

The method of claim 4 wherein the attribute scoring technique specifies weighting to be used when combining the attribute similarity score for that attribute with the attribute similarity score of other attributes.

### Claim 9 (depends on 3)

The method of claim 3 wherein a filter indicates a threshold filter similarity score wherein, when the filter similarity score generated in accordance with the filter meets the threshold filter similarity score, the item definition matches the item description.

### Claim 10 (depends on 9)

The method of claim 9 wherein a filter indicates a maximum number of item definitions for which filter similarity scores that exceed the threshold filter similarity score are to be identified.

### Claim 11 (depends on 1)

The method of claim 1 wherein a rule includes one or more filters, each filter specifying how to identify item definitions that are similar to the item description.

### Claim 12 (depends on 11)

The method of claim 11 wherein a filter specifies a threshold filter similarity score that indicates the threshold similarity score for that filter needed for a match.

### Claim 13 (depends on 11)

The method of claim 11 wherein a filter specifies a maximum number of item definitions that are to be identified as matching based on the filter.

### Claim 14 (depends on 1)

The method of claim 1 including providing a potentially matching threshold similarity score for an item definition to be considered as potentially matching the item description.

### Claim 15 (depends on 1)

The method of claim 1 wherein when more than one item definition has a similarity score that indicates it is a match, identifying those item definitions as ambiguously matching item definitions.

### Claim 16 (depends on 1)

The method of claim 1 including:
generating indexes for attributes, each index for an attribute mapping values of that attribute to item definitions; and
using the indexes to identify candidate item definitions.

### Claim 17 (depends on 16)

The method of claim 16 wherein a rule specifies a criterion for identifying the candidate item definitions.

### Claim 18

A method in a computer system for determining whether an item description matches an item definition in a collection of item definitions, the method comprising:
providing one or more rules for use in determining whether the item description matches an item definition, each rule having one or more filters specifying a scoring metric for calculating a similarity score indicating similarity between an item description and an item definition, wherein a score is assigned to the attributes of the item definition and the item description, and scores of the attributes are aggregated to derive the similarity score;
selecting candidate item definitions that may match the item description;
for each rule,
for each filter in the rule,
calculating similarity scores for candidate item definitions in accordance with the scoring metric of the filter; and
selecting based on the calculated similarity scores the candidate item definition that most closely matches the item description as the matching item definition.

### Claim 19

A computer system for matching item descriptions to item definitions, comprising:
a matching engine subsystem that identifies candidate item definitions for an item description, that calculates similarity scores for candidate item definitions, and that indicates candidate item definitions that match or potentially match an item description based on the similarity scores, wherein a score is assigned to the attributes of the item definition and the item description, and scores of the attributes are aggregated to derive the similarity score; and
a manual matching subsystem that allows a user to designate a potentially matching item definition as matching an item description.

### Claim 20 (depends on 19)

The computer system of claim 19 wherein when only one candidate item definition is indicated as matching the item description, designating that candidate item definition as matching the item description.

### Claim 21 (depends on 19)

The computer system of claim 19 wherein the manual matching subsystem allows the user to designate one of two or more matching item definitions as matching the item description.

### Claim 22 (depends on 19)

The computer system of claim 19 wherein the matching engine subsystem uses rules for calculating the similarity scores.

### Claim 23 (depends on 22)

The computer system of claim 22 wherein a rule includes scoring metrics for calculating similarity scores based on different attributes of an item description and item definitions.

### Claim 24 (depends on 23)

The computer system of claim 23 wherein a scoring metric specifies how to combine similarity scores for different attributes.

### Claim 25 (depends on 24)

The computer system of claim 24 wherein the similarity scores of attributes are assigned weights for their combining.

### Claim 26 (depends on 19)

The computer system of claim 19 wherein the matching engine subsystem uses indexes of attributes into item definitions to identify candidate item definitions.

### Claim 27 (depends on 19)

The computer system of claim 19 wherein an item description specifies a product that a merchant has in inventory and an item definition corresponds to an entry in a catalog of products.

### Claim 28

A computer-readable medium containing instructions for identifying an item definition that matches an item description, by a method comprising:
providing one or more rules that specify how to generate a similarity score based on similarity between the item definition and an item description;
for each of a plurality of item definitions, generating a similarity score for the item definition and the item description in accordance with the rules, wherein a score is assigned to the attributes of the item definition and the item description, and scores of the attributes are aggregated to derive the similarity score; and
selecting the item definition whose generated similarity score indicates it is most similar to the item description as the matching item definition.

### Claim 29 (depends on 28)

The computer-readable medium of claim 28 wherein an item definition and an item description have attributes with values and wherein a rule indicates how to generate a similarity score based on similarity between the values of the attributes of the item definition and the item description.

### Claim 30 (depends on 29)

The computer-readable medium of claim 29 wherein a rule indicates how to generate a similarity score based on similarity of the values of multiple attributes.

### Claim 31 (depends on 28)

The computer-readable medium of claim 28 wherein a rule includes one or more filters, each filter specifying how to generate a filter similarity score, wherein the similarity score for an item definition is the filter similarity score that indicates the item definition is most similar to the item description.

### Claim 32 (depends on 31)

The computer-readable medium of claim 31 wherein an item definition and an item description have attributes with values and wherein a filter specifies an attribute scoring technique for one or more attributes that generates an attribute similarity score for that attribute and specifies how to combine the generated attribute similarity scores to generate a filter similarity score.

### Claim 33 (depends on 32)

The computer-readable medium of claim 32 wherein an attribute scoring technique specifies a minimum score for the attribute, wherein, when the minimum score is not met, the filter similarity score for the filter is set to indicate no similarity.

### Claim 34 (depends on 32)

The computer-readable medium of claim 32 wherein an attribute scoring technique specifies that, when the attribute is not present in either the item definition or the item description, the filter similarity score is set to indicate no similarity.

### Claim 35 (depends on 32)

The computer-readable medium of claim 32 wherein an attribute scoring technique specifies that, when a value for the attribute is not valid in either the item definition or the item description, the filter similarity score is set to indicate no similarity.

### Claim 36 (depends on 32)

The computer-readable medium of claim 32 wherein the attribute scoring technique specifies weighting to be used when combining the attribute similarity score for that attribute with the attribute similarity score of other attributes.

### Claim 37 (depends on 31)

The computer-readable medium of claim 31 wherein a filter indicates a threshold filter similarity score wherein, when the filter similarity score generated in accordance with the filter meets the threshold filter similarity score, the item definition matches the item description.

### Claim 38 (depends on 37)

The computer-readable medium of claim 37 wherein a filter indicates a maximum number of item definitions for which filter similarity scores that exceed the threshold filter similarity score are to be identified.

### Claim 39 (depends on 28)

The computer-readable medium of claim 28 wherein a rule includes one or more filters, each filter specifying how to identify item definitions that are similar to the item description.

### Claim 40 (depends on 39)

The computer-readable medium of claim 39 wherein a filter specifies a threshold filter similarity score that indicates the threshold similarity score for that filter needed for a match.

### Claim 41 (depends on 39)

The computer-readable medium of claim 39 wherein a filter specifies a maximum number of item definitions that are to be identified as matching based on the filter.

### Claim 42 (depends on 28)

The computer-readable medium of claim 28 including providing a potentially matching threshold similarity score for an item definition to be considered as potentially matching the item description.

### Claim 43 (depends on 28)

The computer-readable medium of claim 28 wherein when more than one item definition has a similarity score that indicates it is a match, identifying those item definitions as ambiguously matching item definitions.

### Claim 44 (depends on 28)

The computer-readable medium of claim 28 wherein an item definition and an item description have attributes with values and including:
generating indexes for attributes, each index for an attribute mapping values of that attribute to item definitions; and
using the indexes to identify candidate item definitions.

### Claim 45

A computer system for identifying an item definition that matches an item description, comprising:
means for providing one or more rules that specify how to generate a similarity score based on similarity between the item definition and an item description;
means for generating a similarity score for the item definition and the item description and each of a plurality of item definitions in accordance with the rules, wherein a score is assigned to the attributes of the item definition and the item description, and scores of the attributes are aggregated to derive the similarity score; and
means for selecting the item definition whose generated similarity score indicates it is most similar to the item description as the matching item definition.

### Claim 46 (depends on 45)

The computer system of claim 45 wherein an item definition and an item description have attributes with values and wherein a rule indicates how to generate a similarity score based on similarity between the values of the attributes of the item definition and the item description.

### Claim 47 (depends on 45)

The computer system of claim 45 wherein a rule includes one or more filters, each filter specifying how to generate a filter similarity score, wherein the similarity score for an item definition is the filter similarity score that indicates the item definition is most similar to the item description.

### Claim 48 (depends on 47)

The computer system of claim 47 wherein an item definition and an item description have attributes with values and wherein a filter specifies an attribute scoring technique for one or more attributes that generates an attribute similarity score for that attribute and specifies how to combine the generated attribute similarity scores to generate a filter similarity score.

### Claim 49 (depends on 45)

The computer system of claim 45 wherein a rule includes one or more filters, each filter specifying how to identify item definitions that are similar to the item description.

### Claim 50 (depends on 49)

The computer system of claim 49 wherein a filter specifies a threshold filter similarity score that indicates the threshold similarity score for that filter needed for a match.

### Claim 51 (depends on 49)

The computer system of claim 49 wherein a filter specifies a maximum number of item definitions that are to be identified as matching based on the filter.

### Claim 52 (depends on 45)

The computer system of claim 45 including means for providing a potentially matching threshold similarity score for an item definition to be considered as potentially matching the item description.

### Claim 53 (depends on 45)

The computer system of claim 45 including mean for identifying those item definitions as ambiguously matching item definitions when more than one item definition has a similarity score that indicates it is a match.
