Browse Prior Art Database

Method and System for Advanced Selection of Relevant Advertisements

IP.com Disclosure Number: IPCOM000198738D
Publication Date: 2010-Aug-13
Document File: 4 page(s) / 113K

Publishing Venue

The IP.com Prior Art Database

Related People

Rishit Ratnakar Shetty: INVENTOR [+3]

Abstract

A method and system for advanced selection of relevant advertisements is disclosed. Advertisements are selected even for search queries from users having keywords that are not covered by a set of keywords for which advertisers have bid.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 36% of the total text.

Method and System for Advanced Selection of Relevant Advertisements

Abstract

A method and system for advanced selection of relevant advertisements is disclosed.  Advertisements are selected even for search queries from users having keywords that are not covered by a set of keywords for which advertisers have bid.

Description

Disclosed is a method and system for advanced selection of relevant advertisements.  Advertisements are selected even for search queries from users with keywords which are not covered by a set of keywords for which advertisers have bid.  Such search queries are called non-covered queries.

The method and system disclosed herein, uses multiple algorithms to broaden a scope of search queries from users, while maintaining user intention and selecting relevant advertisements corresponding to the search queries.  Relevant advertisements corresponding to the search queries are served to the users as illustrated in the figure.

Figure

The users provide the search queries using web browsers.  Thereafter, the search queries are sent to a Search System via World Wide Web (WWW).  The search system includes a Search Front End, a Search Back End System, a Click Redirect System, a Sponsored Search Front End, an Advanced Match System and a Computational Grid.  The Search Front End respond to the search queries by requesting algorithmic results from the Search Back End System.  Relevant advertisements are provided by the Sponsored Search Front End.  When a user clicks on an advertisement, the Click Redirect System redirects the user to a destination pointed by URL of the advertisement.  Impression and Click logs containing algorithmic, advertisement impressions and clicks along with the corresponding search queries are collected by the Computational Grid for offline analysis.

The Computational Grid consists of multiple nodes and performs offline processing of impressions and click logs to generate query rewrite mappings and data sets used for advertisement selection process.  The Computational Grid performs Query Rewrites Generations for the search queries using query rewrite technologies such as Modular Optimized Dynamic Sets (MODS), Hubble, Oscar, etc.  Accordingly, the Query Rewrites Generation provides one or more query rewrites for each search query.  The query rewrites for search queries is used along with advertisement impressions and click log data in case keywords in the search queries do not have a corresponding set of relevant advertisements.  Hence, the method and system increases coverage for non-covered queries.  In other words, a replacement query is picked for each non-covered query.  The replacement query passes an experimental threshold of being related to keywords in a corresponding search query and also passes a threshold defined for a combination of parameters based upon historical advertisements and click log data.  As described above, q refers to a query from a user which is a non-covered query and q...