Facebook, Inc.

Nikhil Gupta

Shashikant Khandelwal

Gurpreetsingh Baljeetsingh Sachdev

Apurva Jadhav

Gurpreet Singh Sachdev

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.

Method for relevancy ranking of products in online shopping

In one embodiment, a method includes generating a query based on a seed product offer that describes a product being offered for sale and executing the query, using a search engine that normalizes search terms in the query, on a plurality of product offers to determine candidate offers. The candidate offers are likely associated with the product. The method further includes determining one or more common attribute values across corresponding attributes of the candidate offers. The method additionally includes for each attribute value of each candidate offer, scoring the attribute value based on whether it matches one or more of the common attribute values, and updating a set of normalized attribute values for the product to include the attribute value based on whether the score is greater than a predetermined threshold.

Method, media, and system for product clustering

In one embodiment, a method includes receiving a set of normalized attribute values that define a particular item being offered and a corresponding threshold score for each of the normalized attribute values. The method further includes, for each attribute value of a candidate item listing, determining a score for the attribute value based on whether it matches a corresponding normalized attribute value, and determining whether the score is above or below the corresponding threshold score for the normalized attribute value. The method still further includes determining whether the candidate item listing describes the particular item being offered based on a set of rules that define an acceptable number of scores that are allowed to be below the threshold scores.

Method, media, and system for cluster pruning of item listings

In one embodiment, a method includes receiving a plurality of candidate offers that are likely associated with a product being offered for sale. Each candidate offer is associated with a common set of attributes, wherein at least one of the attributes in the common set uniquely identifies the product being offered for sale. The method further includes, for each attribute value of each candidate offer, scoring the attribute value based on an aggregate number of times that the attribute value is found across corresponding attributes of other candidate offers in the plurality, and adding the scores for a first pair of attributes to determine a cumulative score for the candidate offer. One of the attributes in the first pair is the at least one attribute that uniquely identifies the product. The method still further includes updating a set of normalized attribute values for the product to include the attribute values for the first pair of attributes from a highest scoring candidate offer.

Product scoring for clustering

In one embodiment, a method includes extracting a document object model (DOM) for a content page, wherein the DOM comprises a hierarchical tree-based data structure. The method also includes identifying candidate nodes in the DOM based on a context of the nodes, wherein the candidate nodes may correspond to listing items. The method additionally includes for each of the candidate nodes, locating its parent and child nodes by traversing the DOM from the candidate node, extracting information from the candidate node and its parent and child nodes, and assessing whether the candidate node qualifies as a listing item based on whether the extracted information fulfills a required set of characteristics for a listing item.

Growing community of inventors

Palo Alto, CA, United States of America

Nikhil Gupta