Browse Prior Art Database

Method and System to Validate Reviews/Recommendations of Content in a Social Software System

IP.com Disclosure Number: IPCOM000240862D
Publication Date: 2015-Mar-06
Document File: 2 page(s) / 80K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed are a method and system to make use of a person's social network and the connections between the user’s network and others to generate an affinity score, which is then applied to indicate to the user which online opinions are relevant and credible to the user.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 2

Method and System to Validate Reviews/Recommendations of Content in a Social Software System

In online and enterprise systems for evaluating items (e.g., restaurant reviews, comments on a document, etc.), users can read the reviews of others. However, unless the user happens to personally know one or more of the reviewers, the validity or trustworthiness of the reviewers' comments cannot be further confirmed. Often, reviews can be false, in either the negative or the positive direction of opinion.

An inventory of rating services on the web uses manual processing of reviews, and features such as captcha to eliminate robotic responders only groom reviews. Internal social software relies solely on self-policing to identify and remove bogus reviews and comments, and is inadequate as people do not often take the time to read and engage to make this a more useful, holistic approach.

The novel solution proposes to make use of a person's social network and the connections between the user's network and others to indicate a social "trustworthiness" value. In addition, this method accumulates a "topic affinity" score for the connection between two people. This score indicates, for a specific topic, how much or little an affinity the people share.

As a result, when a user is viewing some content and inspecting the reviews/comments, the system augments that reviewer's rating of the item with the relative affinity score.

When a user looks at content and sees one or more reviews from people directly in a personal or professional network, the system does not need to do anything, as that user already has innately computed the natural human affinity score that person. For example, User A and User B are friends, have similar tastes, and both like Mexican food; therefore,

when User A sees a review of a Mexican restaurant by User B, User A already has the data to measure confidence in User B's review.

However, in wide enterprise social systems, it is unlikely the user would know one or more people who have provided a review. The system is able to compute affinity scores between the user and the list of reviewers to give the user a sense of trustworthiness to the reviews.

To do this, the system inspects a user's social network and computes the social proximity of the list of reviewers to that user's network. Next, it computes the affinity score of people by walking the list of how the user is connected to the reviewer, and accumulates the affinity score as it traverses this list.

Using the example above, User A reads a Mexican restaurant review by User C, who is a friend of User...