Browse Prior Art Database

Auction Mechanism and Resource Economy For Recommendations

IP.com Disclosure Number: IPCOM000238356D
Publication Date: 2014-Aug-19
Document File: 8 page(s) / 27K

Publishing Venue

The IP.com Prior Art Database

Abstract

A recommendation engine, such as a website that recommends websites to a user or a personal assistant included in a smartphone, may have multiple providers which suggest content. The recommendation engine may allow the providers to score their results to indicate the relative values of their results. To prevent the providers from overscoring their results, each provider may have a fixed number of units that they may divide between different results over a fixed period of time, such as a day or a week. The recommendation engine may give different numbers of units to each provider based on how valuable their results are, such as how often the users selects their results.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 28% of the total text.

Page 01 of 8

Auction Mechanism and Resource Economy For Recommendations

ABSTRACT

A recommendation engine, such as a website that recommends websites to a user or a personal assistant included in a smartphone, may have multiple providers which suggest content. The recommendation engine may allow the providers to score their results to indicate the relative values of their results. To prevent the providers from overscoring their results, each provider may have a fixed number of units that they may divide between different results over a fixed period of time, such as a day or a week. The recommendation engine may give different numbers of units to each provider based on how valuable their results are, such as how often the users selects their results.

    Recommendation engines may recommend content, such as websites, to a user. The recommendation engines may receive content from multiple providers. The providers may each derive data or content from, for example, a news website, the user's browser history, the user's email account, or an intelligent personal assistant. The providers may score the content that they provide to the recommendation engine. The recommendation engine may provide recommendations to the user based on the received scores. The providers may have an incentive to overscore their respective content, so that the recommendation engine will present their content to the user.

    Therefore, a recommendation engine is proposed, which limits the number of units that the providers can use to score their results. The providers can divide their units between different results over a fixed period of time, such as a day or a week. The providers will then have an incentive to invest their units on items where they have a high confidence in the score, thereby preventing overscoring. The recommendation engine may give different numbers of units to each provider based on how valuable their results are, such as how often the users selects their results. The recommendation engine may change the numbers of units given to each provider


Page 02 of 8

periodically or in response to certain events, such as a provider's results being selected very frequently by the user.

    FIG. 1 is a diagram of a browser 100 according to an example embodiment. The browser 100 may present a page that recommends websites. The page recommending websites is an example of a user interface for a recommendation engine. Other user interfaces for recommendation engines include voice-activated personal assistants included in smartphones. The browser 100 may include a title field 102, which includes the title of the webpage. The browser 100 may also have an address field 104, which includes an address, which may be in the form of a URL or an IP address of the webpage. The address field 104 may include the address of a website that recommends other websites. The website may recommend other websites based on a recommendation engine.

    When the browser 100 is recommending websites based on...