Natural language product comparison guide synthesizer

• US 20020095411A1
• Filed: 01/16/2001
• Published: 07/18/2002
• Est. Priority Date: 01/16/2001
• Status: Active Grant

• Current Assignee: Cogentex Incorporated
• Original Assignee: Cogentex Incorporated
• Inventors: White, Michael, Korelsky, Tanya, Caldwell, David Edward
• Granted Patent: US 7,418,447 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 40/56 Natural language generation
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0601 Electronic shopping [e-shopping]

Abstract

A Natural Language Product Comparison Guide Synthesizer that lets developers easily create web-based comparison guides for a given type of product. Each comparison guide will solicit users'"'"' product requirements, let them compare different products, and recommend one or more products based on their preferences. The guide uses automatically generated natural language to facilitate the comparison of products, and to convey and explain its recommendations.

Claims

1. A method of creating an automated interactive product comparison guide for providing customers with comparisons between a plurality of products in a class, each product having a plurality of features, comprising the steps of:

a) developing a set of features that apply to the products in the class;

b) developing groupings of these features, which serve to organize them thematically, and which groupings can themselves be viewed as features;

c) developing a scoring function for each feature, which is used to rank different products based on their values of that feature;

d) developing text snippets for each feature, the snippets being phrases to be used when describing or referring to particular product features;

e) developing user profiles, the profiles being collections of values of features that are considered to be suitable for different types of users of the product class, and which can serve to help users of the product guide narrow down their product preferences quickly;

f) developing snippets for each user profile, the snippets being phrases to be used when describing or referring to particular user profiles.

2. The method of claim 1, in which step (c) further comprises the step of testing the scoring function.

3. The method of claim 1, in which step (d) further comprises the step of testing the snippets of the features.

4. The method of claim 1, in which step (e) further comprises the step of testing the user profiles.

5. The method of claim 1, in which step (f) further comprises the step of testing the snippets of the profiles.

6. The method of claim 1, further comprising the step of providing access to the product comparison guide over a computer network.

7. The method of claim 1, further comprising the step of:

g) providing customers with product comparisons over a computer network, using the user profile snippets, feature snippets, scoring functions and feature values.

8. A method of providing customers with product comparisons over a computer network comprising the steps of:

a) creating a product comparison guide for providing customers with comparisons between a plurality of products in a class, each product having a plurality of features, comprising the steps of;

i) developing a set of features that apply to the products in the class;

ii) developing groupings of these features, which serve to organize them thematically, and which groupings can themselves be viewed as features;

iii) developing a scoring function for each feature, which is used to rank different products based on their values of that feature;

iv) developing text snippets for each feature, the snippets being phrases to be used when describing or referring to particular product features;

v) developing user profiles, the profiles being collections of values of features that are considered to be suitable for different types of users of the product class, and which can serve to help users of the product guide narrow down their product preferences quickly;

vi) developing snippets for each user profile, the snippets being phrases to be used when describing or referring to particular user profiles;

b) accepting a request from a customer over a computer network, the request including preferred values for product features;

c) generating a ranked list of products using preferred values entered in step (b), product feature data, and feature scoring functions, such that rankings are based on each product'"'"'s score for each feature as computed using its scoring function, which is then weighted according to the preferred values;

d) combining generic phrases with text snippets for individual features and user profiles to generate a display; and

e) returning the display to the customer over the computer network.

9. The method of claim 8, in which step (b) further comprises the step of storing the preferred values in a user preferences database.

10. The method of claim 8, in which the preferred values from step (b) are requested by specifying a user preference from a database.

11. The method of claim 8, in which the display generated in step (d) comprises at least a display and explanation of product rankings.

12. The method of claim 8, in which the display generated in step (d) comprises at least a display and explanation of a comparison between several products.