Browse Prior Art Database

Method and System for Detecting Comment Spammers by Analyzing Efforts of a User Writing Comments

IP.com Disclosure Number: IPCOM000249788D
Publication Date: 2017-Apr-05
Document File: 2 page(s) / 33K

Publishing Venue

The IP.com Prior Art Database

Related People

Hua Ouyang: INVENTOR [+3]

Abstract

A method and system is disclosed for detecting comment spammers by analyzing efforts of a user writing comments. The method and system utilizes a graph-based approach that quantifies an amount of effort the user exerts over the comments to detect if the user is a comment spammer or not.

This text was extracted from a Microsoft Word document.
This is the abbreviated version, containing approximately 53% of the total text.

Method and System for Detecting Comment Spammers by Analyzing Efforts of a User Writing Comments

Abstract

A method and system is disclosed for detecting comment spammers by analyzing efforts of a user writing comments.  The method and system utilizes a graph-based approach that quantifies an amount of effort the user exerts over the comments to detect if the user is a comment spammer or not.

Description

Typically, a user comments on a post.  The user may be a comment spammer and detecting such comment spammer is challenging.  A comment spam message by a comment spammer is different from other forms of spam in which the comment spam message is usually short, extremely malicious and carefully crafted by humans.  The comment spammer also tends to share IP addresses and browser cookies, especially if the comment spam originates in a botnet.  Even human experts face challenges in differentiating some comment spam messages from legitimate ones.  Relying solely on human experts to detect comment spam is also not feasible.  Human experts simply do not have the bandwidth to deal with the enormous amounts of content generated by the users.  The existing approaches tackling comment spam attempt to classify a comment message as either spam or legitimate.  The comment message is classified as spam or legitimate by considering the characteristics of comments.  The users/senders send the comment spam message separately without considering previous submissions of comments by the same user/sender.  Thus, existing approaches lead to sub-optimal solutions.

Disclosed is a method and system for detecting comment spammers by analyzing efforts of a user writing comments.  The method and system utilizes a graph-based approach that quantifies an amount of effort the user exerts over the comments to detect if the user is a comment spammer or not.

In one implementation, the method and system operates on a bipartite graph of us...