Browse Prior Art Database

Method for Enhancing Social Recommendation

IP.com Disclosure Number: IPCOM000222854D
Publication Date: 2012-Oct-25
Document File: 2 page(s) / 34K

Publishing Venue

The IP.com Prior Art Database

Abstract

A method for enhancing social recommendation of a content is disclosed. When a user wishes to recommend the content to other users on a social network, each of a first set of users who may like the content and a second set of users who may dislike the content are predicted and presented to the user. Based on the user’s selection of one or more users for recommending the content, the method and system may identify patterns in the user’s selection that may indicate, for example, astroturfing.

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 for Enhancing Social Recommendation

Disclosed is a method for enhancing recommendation of a content on a social network. The content may include for example, but is not limited to an image, a video clip, article blog post and the like. The method includes predicting a first set of users of the social network who are most probable to provide positive feedback on the content. The predicting is based on analyzing a history of feedbacks received from users of the social network. Additionally, the method also includes predicting a second set of users of the social network who are least probable to provide positive feedback on the content based on a history of feedbacks received from users of the social network. Subsequently, each of the first set of users and the second set of users are presented to a user when the user wishes to recommend the content to other users of the social network. The presentation may be performed near a user-interface element, for example a "Like" button, meant for receiving feedback. The user may then make a selection of one or more users from one or more of the first set of users and the second set of users. The method records the selection and tracks such selections by the user over a period of time in order to identify patterns that may indicate, for example, astroturfing.

In an implementation of the method disclosed herein, the content may be recommended by receiving positive feedback from the user through a "Like" button of Facebook*. The "Like" button may be related to the content, which may be for example, embedded in a web page. The method analyzes the corpus of "Likes" in the social network in order to predict each of the first set of users and the second set of users.

In an embodiment, the prediction of each of the first set of users and the second set of users may be performed by using random rainforest with existing data sets, and training sets. To this end, a series of rules from the trees in the forest may be generated. Further, in order to fine tune the prediction the method may include a few months of training period. Further, the prediction may be scoped based on one or more of time, day, week, month and year. In some embodiments, the method may use negation to find the second set of users based on a prediction of the first set of users. In some other embodiments...