Browse Prior Art Database

Categorical Meta-reviews for Improved Recommendations

IP.com Disclosure Number: IPCOM000235692D
Publication Date: 2014-Mar-20
Document File: 4 page(s) / 36K

Publishing Venue

The IP.com Prior Art Database

Abstract

The disclosed invention describes a system and method to create personalized product recommendations based on analytics applied to a new category of meta-reviews associated with product reviews.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 53% of the total text.

Page 01 of 4

Categorical Meta-reviews for Improved Recommendations

User ratings at online stores provide a wealth of information, but some ratings are often more applicable than others. For example, a video game may have high ratings from expert video game players and poor reviews from casual players.

Shoppers should have an indication of how valuable a review will be to them, or only be shown applicable reviews. Furthermore, users should not be required to enter personal details to receive custom-tailored recommendations. The proposed categorical recommendation engine analyzes user ratings of reviews to predict the applicability of other reviews. These weighted reviews are then used to provide customized product feedback to the user.

High-level Steps

1. When a review is entered, collect the reviewer's background details for categories that relate to the product.

2. Other users rate how applicable that review is to them (highly applicable to not applicable).

3. Analytics engine measures an individual user's ratings to determine how categories affect that user.


4. Analytics engine predicts applicability of other reviews to that user.

5. Store tailors the product reviews and aggregate scores based on results from analytics engine.

Detailed Steps

1. The store places products into product areas (electronics, clothing, etc.) and when reviews are entered for products, a set of reviewer demographic categories are collected based on the product area. For example someone reviewing a new video game on an online store would enter four categories that could affect impressions of a video game: age, time, experience and platform.

Review: "This my friend, is an award winning game for any strategy gamer. Novice or expert, this is a great buy for your hard-earned money!"

2. Shoppers can then rate any review to indicate how valuable the review was to them. For example, a user could mark they found the above review to be highly valuable, or not at all useful. This rating is different than the typical "73 of 81 people found the following review helpful" because it is collecting the applicability of a review instead of the quality

1


Page 02 of 4

of the review. The score will not shared with other users and is instead used as a data

point to find other reviews that the shopper would find useful.

User rating a review:

3. Based on the results of multiple ratings on reviews, a statistical engine determines the categories that are important to the user and how they...