Method for relevancy ranking of products in online shopping

Patent metadata

Field	Value
Publication number	US-9779441-B1
Application number	US-201313936106-A
Country	US
Kind code	B1
Filing date	Jul 5, 2013
Priority date	Aug 4, 2006
Publication date	Oct 3, 2017
Grant date	Oct 3, 2017

Abstract

Systems and methods for ranking one or more products in online shopping. One or more products are identified based on a search query received from user. The one or more products are ranked based on terms present in the search query. Each of the one or more products has one or more attributes associated with it. An attribute score for each of the one or more products is determined. Further, based on the attribute score, the relevancy of the one or more products is determined. Based on the relevancy, a marginal relevancy score for each of the one or more products is determined. The one or more products are re-ranked based on the marginal relevancy score. The rank of the one or more products can also be modified to optimize revenue generation.

First claim

What is claimed is:

1. A method for ranking one or more products in online shopping, the method comprising the steps of:
extracting a first list of the one or more products based on at least one term in a search query received from a user, the one or more products being ranked in the first list based on a term score associated with each of the one or more products;
identifying a plurality of attributes associated with the one or more products, the plurality of attributes comprising an independent attribute and a dependent attribute, each of the independent and dependent attribute corresponding to one of a brand name, a store name, a style, or a price;
for each of the one or more products:
determining an independent attribute score for the independent attribute based on one or more features associated with the product;
determining a dependent attribute score for the dependent attribute based on the determined independent attribute score and one or more features associated with the product;
determining a relevance score based on the term score and the dependent attribute score associated with the dependent attribute;
determining a marginal relevance score to be associated with the product by:
identifying a ranking of the product in the extracted first list of the one or more ranked products;
identifying a higher ranked product in the extracted first list;
comparing the dependent attribute of the product with the dependent attribute of the higher ranked product;
determining a degree of similarity between the dependent attribute of the product and the higher ranked product based on the comparison;
modifying the relevance score of the product by a factor representing the determined degree of similarity; and

ranking the one or more products based on the marginal relevance score associated with each of the one or more products to generate a second list.

2. The method of claim 1 further comprising the step of determining the term score based on an occurrence of the at least one term in description of the one or more products.

3. The method according to claim 1 , wherein one of the independent attribute or one or more dependent attributes is a brand attribute, and wherein the one or more features associated with the brand attribute is at least one of quality of the one or more products of the brand name, the range of designs available in the brand, and the number of merchants selling the brand.

4. The method according to claim 1 , wherein one of the independent attribute or one or more dependent attributes is a store attribute, and wherein the one or more features associated with the store attribute is at least one of a network traffic ranking of a store, number of users querying for the store, number of click-outs by users on the one or more products sold by the store, financials of the store, payment methods accepted by the store, website security and trust ranking of the store, and local presence of the store.

5. The method according to claim 1 , wherein the one or more features associated with the independent or one or more dependent attributes are associated with market demand-supply characteristics for the one or more products, wherein the one or more features associated with the market demand-supply characteristics for the one or more products are at least one of quality of the one or more products of a brand, the range of designs available in the brand, the number of merchants selling products by the brand and quality and financials of the merchants selling the products.

6. The method of claim 1 , wherein the relevance score is determined using an equation expressed as:
Relevance score( r )=term score( r )*Attribute score( r ),
r being a product from the one or more products.

7. The method of claim 6 further comprising initializing the marginal relevance score as the relevance score.

8. The method of claim 7 , wherein the marginal relevance score is modified using an equation expressed as:
Marginal relevance( r ′)=Relevance score( r ′)*(α^similarity( r,r ′)),

a being a predetermined constant, r being an array of products ranked above the product r′, and similarity(r, r′) being a measure of similarity between the r′ and r.

9. A method for ranking one or more products in online shopping, the method comprising the steps of:
receiving a first ranked list of one or more products each associated with one or more attributes, the one or more products being ranked based on a term score associated with each of the one or more products to a search query received from a user;
identifying a plurality of attributes associated with the one or more products, the plurality of attributes comprising an independent attribute and a dependent attribute, each of the independent and dependent attribute corresponding to one of a brand name, a store name, a style, or a price;
for each of the one or more products:
determining an independent attribute score for the independent attribute based on one or more features associated with the product;
determining a dependent attribute score for the dependent attribute based on the determined independent attribute score and one or more features associated with the product;
determining a relevance score based on the term score and the dependent attribute score associated with the dependent attribute;
determining a marginal relevance score to be associated with the product by:
identifying a ranking of the product in the extracted first list of the one or more ranked products;
identifying a higher ranked product in the extracted first list;
comparing the dependent attribute of the product with the dependent attribute of the higher ranked product;
determining a degree of similarity between the dependent attribute of the product and the higher ranked product based on the comparison;
modifying the relevance score of the product by a factor representing the determined degree of similarity;

ranking the one or more products based on the marginal relevance score associated with each of the one or more products to generate a second ranked list;
determining a discounted cumulative gain (DCG) score for second ranked list based on the marginal relevance score associated with each of the one or more products in the second ranked list; and
modifying ranking of the one or more products in the second ranked list, wherein modifying the ranking includes:
determining a position adjustment for at least one product of the one or more products in the second ranked list based at least in part on a cost per click (CPC) associated with each of the one or more products;
determining a loss of DCG score incurred by the second ranked list due to adjusting the position of the at least one product; and
responsive to determining that the loss of DCG score is less than a predetermined threshold value, modifying the ranking of the one or more products in the second list.

10. The method of claim 9 further comprising the step of extracting the first ranked list of the one or more products based on at least one term in the search query, wherein the one or more products are ranked in the first ranked list based on a term score associated with each of the one or more products, wherein the term score is indicative of relevancy.

11. The method of claim 10 further comprising the step of determining the term score based on an occurrence of the at least one term in description of the one or more products.

12. The method of claim 9 wherein determining a position adjustment for at least one product of the one or more products in the second ranked list based at least in part on a cost per click (CPC) associated with each of the one or more products compris

Assignees

• Facebook Inc

Inventors

• Jadhav Apurva
• Khandelwal Shashikant
• Gupta Nikhil
• Sachdev Gurpreet Singh

Classifications

• G06Q30/0282

  Rating or review of business operators or products · CPC title

• G06F17/3053

  Physics · mapped topic

• G06Q30/0627 Primary

  by specifying product or service characteristics, e.g. product dimensions · CPC title

• G06F16/838

  Presentation of query results · CPC title

• G06F16/338

  Presentation of query results · CPC title