Browse Prior Art Database

An alternative method of weighting user ratings

IP.com Disclosure Number: IPCOM000237838D
Publication Date: 2014-Jul-16
Document File: 2 page(s) / 51K

Publishing Venue

The IP.com Prior Art Database

Abstract

Review systems are commonplace across the internet for many different fields including, for example, retail and entertainment. These systems can be subject to bias. This article describes a novel way of accounting for this bias to provide better reviews.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 54% of the total text.

Page 01 of 2

An alternative method of weighting user ratings

Review systems are commonplace across the internet for many different fields including, for example, retail and entertainment. Typically user ratings are averaged to provide an overall rating for a product, service, or other thing being reviewed. The disadvantage of this approach is that people have different expectations and tolerance so it can be difficult to accurately compare multiple ratings from different users. This problem is sometimes addressed through the use of guidelines in the rating system, for example, hints that a rating of 1 would be that the product didn't work at all and a rating of 5 that it is the best product ever bought. The drawback of this guidance is that it is still subject to user interpretation. Another approach to solving this problem is to calculate a weighted rating of some kind. A popular film review website, for example, uses a true Bayesian estimate from its user ratings to calculate the top 250 rated films, which accounts for the average rating across the whole catalogue and the number of voters for example. This approach addresses some of the difficulties of variation from multiple users by using a global mean, but it cannot account for individual user bias without a large number of ratings.

    The idea herein is to eliminate some user specific bias from item ratings and quickly identify the best and worst items in a catalogue. This is achieved by maintaining a user average rating...